By Trushil Shah, MD, MSc

Assistant Professor of Medicine, University of Texas Southwestern Medical Center, Dallas

Dr. Shah reports he receives grant/research support from Actelion Pharmaceuticals, Liquidia Technologies, and Bayer Pharmaceuticals, and is a consultant for and on the speakers bureau of Gilead Sciences.

The first use of oxygen as a therapy dates to 1885 when it was used to treat pneumonia.1 Since then, use of oxygen therapy has become one of the most common treatments in hospitalized patients. Hypoxia has been independently associated with mortality across various diseases, and it is common knowledge that treatment of hypoxia is critical for survival. While hypoxia can result in several adverse outcomes and oxygen therapy is warranted to achieve normoxia, data from multiple studies show that a large proportion of patients receive oxygen therapy in the absence of this indication.2 Be it in an ambulance, ED, medicine floor, or ICU, many patients receiving oxygen therapy do not have documented hypoxemia. At times, oxygen is administered even without a physician prescription.3,4 Oxygen use has become ubiquitous to medical practice.

More than 25% of all ED patients, as well as most stroke and myocardial infarction patients, receive oxygen therapy.3,5 An audit of oxygen use in a Brooklyn state hospital revealed only 19% of patients on supplemental oxygen had a clear indication, 53% had no active order for supplementation, and 57% were not on continuous bedside pulse oximetry monitoring despite supplemental oxygen.6 With such ubiquitous use of oxygen, the medical community and patients assume there is no harm, and perhaps even potential benefit, associated with its use.7 Current guidelines for using oxygen therapy in medically ill patients are inconsistent and lack consensus on a safe upper limit for oxygenation. Because of the sigmoid nature of the oxygen-hemoglobin dissociation curve, at higher SpO2 readings, there is an exponential increase in PaO2.

In contrast, oxygen toxicity has been studied since the 1950s. Many animal studies have revealed different mechanisms of damage.8-10 Hyperoxia happens when high amounts of reactive oxygen species (ROS) overwhelm natural antioxidant defenses, leading to cell death and apoptosis. The increase in ROS accelerates the release of endogenous damage-associated molecular pattern molecules that stimulate an inflammatory response, especially in the lungs, and cause vasoconstriction, likely because of reduced nitric oxide levels.11,12 The lung is a particularly susceptible target; hyperoxia can cause acute lung injury. Hyperoxia-induced acute lung injury (HALI) is associated with alteration in surfactant protein composition, decreased mucociliary clearance, and cellular damage resulting in atelectasis, a reduction in lung compliance, and increased susceptibility to infection.13 Hyperoxemia-induced vasoconstriction can lead to a reduction in coronary blood flow, decrease cardiac output, and alter microvascular perfusion, too.11,13 The severity of HALI is directly proportional both to the PaO2 (particularly above a rate of 450 mmHg or an FiO2 of 0.6) and exposure time.14

Over the last decade, more clinical studies have shown adverse effects of hyperoxia in different patient populations and its association with increased mortality.15-17 In a meta-analysis, Chu et al synthesized data from 25 randomized, controlled trials comparing a liberal oxygen approach to a conservative approach. They included 16,037 patients with sepsis, critical illness, stroke, trauma, myocardial infarction, cardiac arrest, and emergency surgery. The authors found that liberal oxygen therapy was associated with increased in-hospital mortality, 30-day mortality, and mortality at longest follow-up. The following sections include more details about specific subgroups relevant to ICU practice and a review of the current data on oxygen therapy in these patients.

Myocardial Infarction

Since the early 1900s, it has been routine practice to provide oxygen supplementation to patients with ST-elevation myocardial infarction (STEMI), regardless of their baseline SpO2.18 More recently, accumulating evidence suggests that hyperoxia actually may be harmful in myocardial infarction patients. The authors of the AVOID trial compared 8 L oxygen to room air in 441 patients with STEMI without hypoxia. They found an increase in myocardial infarct size in the oxygen therapy group at six months and no benefit.19 Recently, in the DETO2X-AMI trial, which included 6,629 patients, showed no benefit regarding supplemental oxygen in patients without hypoxemia.20 Abuzaid et al further confirmed this in a meta-analysis of six randomized, controlled trials.21 Based on current data, supplemental oxygen should be used only in patients with myocardial infarction with baseline hypoxemia to a goal of SpO2 between 90% and 95%, remembering that hyperoxia can be harmful.22

Cardiac Arrest

Current guidelines support the usual practice of giving 100% FiO2 in the setting of cardiac arrest and immediately after achieving return of spontaneous circulation (ROSC).23 Two retrospective observational studies revealed that hyperoxia (PaO2 higher than 300 mmHg) during CPR is associated with higher rates of ROSC, lower mortality, and intact neurological survival.16,24 However, this may not be a function of the administered amount of FiO2, but could represent better native lung function, superior resuscitation quality, and lower illness severity.24 In the absence of data to use lower FiO2 concentrations intra-arrest, it is reasonable to continue to use 100% FiO2 during CPR.

However, after ROSC is achieved, hyperoxia is associated with a higher risk of mortality.16 In a recent meta-analysis of observational studies of in-hospital and out-of-hospital cardiac arrests, Patel et al confirmed this association.16 In a Dutch registry study, Helmerhorst et al showed that PaO2 values in the first 24 hours after cardiac arrest are related to mortality in a U-shape, where both hypoxia and hyperoxia may be harmful.25

Septic Shock and Critically Ill Patients

In an observational cohort study of 14,441 Dutch ICU patients, Helmerhorst and other colleagues found that severe hyperoxia as defined by PaO2 > 200 mmHg was associated with increased mortality and fewer ventilator-free days.26 Moreover, they identified a dose-response relationship of hyperoxia with mortality in the first 24 hours and beyond, with more time spent in hyperoxia associated with increased mortality.26 A recent meta-analysis of two randomized, controlled trials and seven cohort studies in ICU patients revealed that hyperoxia was associated with increased hospital mortality (hazard ratio, 1.58; 95% confidence interval, 1.26-2.0).27 In a randomized, controlled trial that included 442 septic shock patients, Asfar et al compared hyperoxia with 100% FiO2 to normoxia with SpO2 88-95%. Investigators discovered that the hyperoxia group trended toward an increase in mortality, especially in patients with lactate > 2 mmol/L.28,29 The hyperoxia group also experienced a significant increase in serious adverse events, mainly driven by a doubling of ICU-acquired weakness and atelectasis.28


Hypoxemia is associated with worse outcomes in ischemic stroke. Oxygen supplementation may improve outcomes by preventing hypoxemia and secondary brain damage.30 However, hyperoxia is associated with cerebral vasoconstriction, resulting in decreased cerebral blood flow.31 In a large multicenter, cohort study that included 2,894 patients, Rincon et al found that in ventilated stroke patients admitted to the ICU, arterial hyperoxia (PaO2 > 300 mmHg) was associated independently with in-hospital death compared with normoxia or hypoxia.32 Study limitations included its observational approach, the authors not accounting for ventilator-specific data, and the authors not adhering to common endpoints used in neurological outcomes research.32

In a large randomized, controlled trial that included 8,003 patients with acute stroke randomized to continuous low-dose oxygen vs. nocturnal oxygen and control, Roffe et al observed that low-dose oxygen did not improve outcomes of death and disability at three months.33 A recent study of short burst high-flow oxygen (45 L/min) ended early because of excess mortality in the actively treated group. The authors of an ongoing randomized, controlled trial (PROOF) are assessing the use of high-flow oxygen at 40 L/min to maintain viability of ischemic penumbra to allow for a broader window for thrombolysis.34 Current guidelines from the American Heart Association suggest using supplemental oxygen in acute ischemic stroke to maintain SpO2 > 94%.35

Surgical Patients

Supplemental oxygen has been used in surgical patients intra- and postoperatively to decrease the incidence of surgical wound infections.36 The oxidative killing of neutrophils depends on PO2; hence, supplemental oxygen theoretically enhances the bactericidal effects of neutrophils.37 To date, several randomized, controlled trials have been performed in different surgical patient populations comparing hyperoxia to normoxia, and results have been conflicting.36 A meta-analysis of these trials has shown a lower incidence of surgical site infections, but the quality of evidence is low, as many of these trials are prone to bias.36 The authors of a long-term follow-up of the PROXI randomized, controlled trial observed that patients undergoing cancer surgery demonstrated higher mortality rates with high inspired FiO2 (80% vs. 30%).38

Target Oxygen Levels

As mounting evidence shows hyperoxia can be harmful, an important question arises: What is a safe level or range of oxygenation in hospitalized and critically ill patients? To make matters more complex, different targets may be indicated for different subsets of patients (Table 1). Guidelines for supplemental oxygen have been inconsistent across countries and even across specialties. Chu et al’s meta-analysis of 25 randomized, control trials that included 16,037 patients revealed that liberal oxygen therapy increased mortality, with one excess death for an average of 71 patients treated with liberal oxygen therapy.15 Across the trials included in this study, the baseline median SpO2 in the liberal oxygen arm was 96% (range, 94-99%). When this group was exposed to liberal oxygenation, researchers observed an increase in mortality risk that was dose-dependent on the magnitude of increase in SpO2. The results of the ICU-ROX randomized, controlled trial may shed more light on this question.39 However, new evidence and guidelines may not change practice quickly, which will require efforts by physicians, nursing staff, respiratory therapists, and even policymakers.40 Barriers to appropriate oxygen prescription, monitoring, and administration will need to be identified at individual hospital levels and addressed.4,40


Oxygen is not a harmless “drug.” Liberal oxygen therapy is associated with increased harm and mortality across different subpopulations in the ICU. Oxygen supplementation should be reserved only for hypoxic patients (SpO2 < 90%), with a goal SpO2 of < 96%.41 Future studies are needed to establish a specific safe range of oxygenation.


  1. Shultz SM, Hartmann PM. George E Holtzapple (1862-1946) and oxygen therapy for lobar pneumonia: The first reported case (1887) and a review of the contemporary literature to 1899. J Med Biogr 2005;13:201-206.
  2. Albin RJ, et al. Pattern of non-ICU inpatient supplemental oxygen utilization in a university hospital. Chest 1992;102:1672-1675.
  3. Hale KE, et al. Audit of oxygen use in emergency ambulances and in a hospital emergency department. Emerg Med J 2008;25:773-776.
  4. O’Driscoll BR. British Thoracic Society Oxygen Guidelines: Another clinical brick in the wall. Thorax 2017;72:498-499.
  5. Burls A, et al. Oxygen use in acute myocardial infarction: An online survey of health professionals’ practice and beliefs. Emerg Med J 2010;27:283-286.
  6. Nath S, et al. An audit of supplemental oxygen prescribing practices in an inpatient setting and its financial burden. European Respiratory Journal 2018;52:PA3160.
  7. Kelly CA, et al. A wolf in sheep’s clothing? Patients’ and healthcare professionals’ perceptions of oxygen therapy: An interpretative phenomenological analysis. Clin Respir J 2018;12:616-632.
  8. Garner WL, et al. The effects of hyperoxia during fulminant sepsis. Surgery 1989;105:747-751.
  9. Bhandari V, et al. Increased hyperoxia-induced mortality and acute lung injury in IL-13 null mice. J Immunol 2007;178:4993-5000.
  10. Bhandari V, et al. Hyperoxia causes angiopoietin 2-mediated acute lung injury and necrotic cell death. Nat Med 2006;12:1286-1293.
  11. Vincent JL, et al. Harmful effects of hyperoxia in postcardiac arrest, sepsis, traumatic brain injury, or stroke: The importance of individualized oxygen therapy in critically ill patients. Can Respir J 2017;2017:2834956. doi: 10.1155/2017/2834956.
  12. Sjoberg F, Singer M. The medical use of oxygen: A time for critical reappraisal. J Intern Med 2013; 274:505-528.
  13. Damiani E, et al. Oxygen in the critically ill: Friend or foe? Curr Opin Anaesthesiol 2018;31:129-135.
  14. Kallet RH, Matthay MA. Hyperoxic acute lung injury. Respir Care 2013;58:123-141.
  15. Chu DK, et al. Mortality and morbidity in acutely ill adults treated with liberal versus conservative oxygen therapy (IOTA): A systematic review and meta-analysis. Lancet 2018;391:1693-1705.
  16. Patel JK, et al. Association between intra- and post-arrest hyperoxia on mortality in adults with cardiac arrest: A systematic review and meta-analysis. Resuscitation 2018;127:83-88.
  17. Stolmeijer R, et al. A systematic review of the effects of hyperoxia in acutely ill patients: Should we aim for less? Biomed Res Int 2018; May 14;2018:7841295. doi: 10.1155/2018/7841295. eCollection 2018.
  18. Beasley R, et al. Oxygen therapy in myocardial infarction: An historical perspective. J R Soc Med 2007;100:130-133.
  19. Stub D, et al. Air versus oxygen in ST-segment-elevation myocardial infarction. Circulation 2015;131:2143-2150.
  20. Hofmann R, et al. Oxygen therapy in suspected acute myocardial infarction. N Engl J Med 2017;377:1240-1249.
  21. Abuzaid A, et al. Oxygen therapy in patients with acute myocardial infarction: A systemic review and meta-analysis. Am J Med 2018;131:693-701.
  22. Ibanez B, et al. 2017 ESC Guidelines for the management of acute myocardial infarction in patients presenting with ST-segment elevation. The task force for the management of acute myocardial infarction in patients presenting with ST-segment elevation of the European Society of Cardiology (ESC). Eur Heart J 2018;39:119-177.
  23. Field JM, et al. Part 1: Executive summary: 2010 American Heart Association Guidelines for cardiopulmonary resuscitation and emergency cardiovascular care. Circulation 2010;122:S640-S656.
  24. Patel JK, et al. Association of arterial oxygen tension during in-hospital cardiac arrest with return of spontaneous circulation and survival. J Intensive Care Med 2018;33:407-414.
  25. Helmerhorst HJ, et al. Associations of arterial carbon dioxide and arterial oxygen concentrations with hospital mortality after resuscitation from cardiac arrest. Crit Care 2015;19:348.
  26. Helmerhorst HJ, et al. Metrics of arterial hyperoxia and associated outcomes in critical care. Crit Care Med 2017;45:187-195.
  27. You J, et al. Association between arterial hyperoxia and mortality in critically ill patients: A systematic review and meta-analysis. J Crit Care 2018;47:260-268.
  28. Asfar P, et al. Hyperoxia and hypertonic saline in patients with septic shock (HYPERS2S): A two-by-two factorial, multicentre, randomised, clinical trial. Lancet Respir Med 2017;5:180-190.
  29. Demiselle J, et al. Hyperoxia toxicity in septic shock patients according to the Sepsis-3 criteria: A post hoc analysis of the HYPER2S trial. Ann Intensive Care 2018;8:90.
  30. Rowat AM, et al. Hypoxaemia in acute stroke is frequent and worsens outcome. Cerebrovasc Dis 2006;21:166-172.
  31. Floyd TF, et al. Independent cerebral vasoconstrictive effects of hyperoxia and accompanying arterial hypocapnia at 1 ATA. J Appl Physiol (1985) 2003;95:2453-2461.
  32. Rincon F, et al. Association between hyperoxia and mortality after stroke: A multicenter cohort study. Crit Care Med 2014;42:387-396.
  33. Roffe C, et al. Effect of routine low-dose oxygen supplementation on death and disability in adults with acute stroke: The stroke oxygen study randomized clinical trial. JAMA 2017;318:1125-1135.
  34. Poli S, et al. Abstract TMP15: Penumbral Rescue by Normobaric O2 in Ischemic Stroke With Target Mismatch ProFile (PROOF). Stroke 2018;49:ATMP15.
  35. Powers WJ, et al. 2018 guidelines for the early management of patients with acute ischemic stroke: A guideline for healthcare professionals from the American Heart Association/American Stroke Association. Stroke 2018;49:e46-e110.
  36. Wetterslev J, et al. The effects of high perioperative inspiratory oxygen fraction for adult surgical patients. Cochrane Database Syst Rev 2015; Jun 25;(6):CD008884. doi: 10.1002/14651858.CD008884.pub2.
  37. Allen DB, et al. Wound hypoxia and acidosis limit neutrophil bacterial killing mechanisms. Arch Surg 1997;132:991-996.
  38. Meyhoff CS, et al. Increased long-term mortality after a high perioperative inspiratory oxygen fraction during abdominal surgery: Follow-up of a randomized clinical trial. Anesth Analg 2012;115:849-854.
  39. Young PJ, et al. Intensive care unit randomised trial comparing two approaches to oxygen therapy (ICU-ROX): Results of the pilot phase. Crit Care Resusc 2017;19:344-354.
  40. O’Driscoll BR, et al. BTS guideline for oxygen use in adults in healthcare and emergency settings. Thorax 2017;72:ii1-ii90.
  41. Siemieniuk RAC, et al. Oxygen therapy for acutely ill medical patients: A clinical practice guideline. BMJ 2018;363: k4169.

Table 1. Level of Evidence and Recommended SpO2 Range in Different Subsets of Critically Ill Patients

Patient Population

Level of Evidence

Recommended SpO2


Myocardial infarction



Hyperoxia increases infarct size

Cardiac arrest

Medium to low


100% FiO2 recommended during resuscitation
of cardiac arrest; SpO
2 goal listed to be used
after return of spontaneous circulation

Mixed medical ICU patients



Hyperoxia may increase ICU-acquired weakness
and atelectasis




Hyperoxia is not beneficial for either ischemic
or hemorrhagic stroke, but its effect on ischemic
penumbra is unknown

Postoperative surgical



Hyperoxia may decrease surgical site infections
but needs to be balanced against other risks
of hyperoxia