By Samuel Nadler, MD, PhD
Critical Care, Pulmonary Medicine, The Polyclinic Madison Center, Seattle; Clinical Instructor, University of Washington, Seattle
Providing supplemental oxygen to patients with hypoxemic respiratory failure is standard practice in the intensive care unit (ICU). Yet, in the first publication describing the acute respiratory distress syndrome (ARDS), Ashbaugh and colleagues responded to concerns regarding the delivery of excessive oxygen: “None of our patients had received high concentrations of oxygen for prolonged periods before the onset of respiratory distress. Oxygen toxicity is, therefore, unlikely to have been an aetiologic agent in this series.”1 Therapeutic targets for supplemental oxygen delivery remain controversial to this day. Trials of liberal vs. conservative oxygen therapy were last reviewed in this journal in 2019.2 With multiple new trials recently published, this feature serves to update the evidence regarding therapeutic oxygen delivery in hypoxemic respiratory failure.
Notable Studies
Several large clinical trials raised concerns for oxygen toxicity in patients receiving supplemental oxygen. In 2016, the Oxygen-ICU trial presented evidence that conservative oxygen therapy compared with conventional therapy led to lower ICU mortality.3 This was a single-center, open-label study of 434 patients. Conservative therapy was defined as the lowest amount of oxygen delivered to achieve an arterial partial pressure of oxygen (PaO2) of 70 mmHg to 100 mmHg or oxygen saturation (SpO2) of 94% to 98%, while the conventional group received a fraction of inspired oxygen (FiO2) > 0.4, allowing PaO2 up to 150 mmHg and SpO2 97% to 100%. The study was stopped early because of an earthquake at the study center. The unplanned interim analysis showed lower ICU mortality in the conservative group as compared with conventional therapy (11.6% vs. 20.2%, P = 0.01).3 The study population was notable for a high percentage of surgical patients in each group (64.3% and 60.7%, respectively), although most had respiratory failure (56.0% and 59.2%, respectively) or were mechanically ventilated (66.2% and 67.9%, respectively).3 Interestingly, the authors noted a U-shaped association between time-weighted PaO2 values and mortality, suggesting both high and low PaO2 levels were injurious. Increased mortality with conventional oxygen delivery was postulated as a result of increased rates of bacteremia and shock from suppression of the innate immune system.
Similarly, the Improving Oxygen Therapy in Acute illness (IOTA) systematic review reported liberal oxygen therapy increased mortality.4 This meta-analysis included studies of liberal vs. conservative oxygenation strategies in adult patients admitted to the ICU prior to 2017. A total of 25 studies covering 16,037 patients admitted with critical illness, stroke, myocardial infarction, and emergency surgery were included. It is important to note that 12 of the 25 studies excluded patients with baseline hypoxemia, and all 25 studies excluded patients with severe baseline hypoxemia as defined by PaO2/FiO2 < 100. Overall, a liberal oxygenation strategy was associated with increased mortality compared with a conservative approach (hazard ratio [HR], 1.11; 95% confidence interval [CI], 1-1.24).4 Unlike Oxygen-ICU, the risk of hospital-acquired infections in the analysis was not statistically different between the groups, although patients admitted for emergency surgery did have lower infection rates with the conservative strategy.
More relevant to hypoxemic respiratory failure, Aggarwal et al reviewed 10 clinical trials within the ARDS Network hospitals between 1996 and 2013.5 They defined excess oxygen exposure as any value of FiO2 > 0.5 in a patient with PaO2 > 80 mmHg. Among the 2,994 patients who met ARDS criteria, in-hospital mortality increased with excess oxygen delivery in a dose-dependent manner (interquartile odds ratio [OR], 1.20; 95% CI, 1.11-1.31).5 In the multivariate analysis, the effect of excess oxygen on in-hospital mortality was less than the contribution of age (OR, 2.01; 95% CI, 1.75-2.31) or APACHE III score (OR, 2.75; 95% CI, 2.41-3.15).5
The HYPERS2S study randomized 234 patients who were mechanically ventilated with septic shock to either hyperoxia (FiO2 = 1.0) or normoxia (SpO2 88% to 95%) as well as isotonic vs. hypertonic saline administration.6 This study was stopped early because of safety concerns, but suggested higher 90-day mortality in the hyperoxia group compared to the normoxia group (48% vs. 41%, P = 0.16).6 In contrast, the CLOSE trial examined the feasibility of a randomized controlled trial of different oxygenation targets in mechanically ventilated ICU patients and showed no difference in 90-day mortality in conservative vs. liberal oxygenation groups (40% vs. 37%, P = 0.75).7
Mechanisms of Oxygen Toxicity
Considering the evidence that excess delivery of oxygen might increase mortality, what should an appropriate target be? Remarkably, no safe level of supplemental oxygen delivery has been established in humans.8 Indeed, hyperoxia has been hypothesized as the second hit that promotes the development of ARDS after an initial injury. While hyperoxia has not been a selective force during evolution, humans have well-characterized biologic responses to both hypoxia and hyperoxia.9 Reactive oxygen species (ROS) generated by normal mitochondrial activity are detoxified by cellular defenses, such as superoxide dismutase, glutathione S-transferase, catalase, and NADPH quinone oxidoreductase. Exposure to hyperoxia can increase ROS production. If this production exceeds cellular responses to oxidative stress, cellular injury can occur. Even in healthy, non-smoking adults without evidence of lung injury, exposure to FiO2 of 0.5 led to increased alveolar neutrophils and oxidative damage evidenced by increased lipid peroxidation and leukotriene release.10 Furthermore, supra-normal PaO2 inhibits hypoxemic pulmonary vasoconstriction, leading to absorption atelectasis, low ventilation/perfusion ratio, and increased shunt.10 In disease states, there is evidence that hyperoxia worsens outcomes with stroke, sepsis, myocardial ischemia, and emergency surgery.2,4
Reconsidering the EFFECT of Higher FIO2 in Hypoxemic Respiratory Failure
More recently, several large studies have called into question the concerns regarding liberal oxygen therapy. The ICU-ROX study randomized 1,000 patients who were expected to require mechanical ventilation to conservative or usual oxygen therapy.11 The conservative protocol decreased or eliminated supplemental oxygen delivery to maintain SpO2 < 97% while assuring SpO2 > 91%. In the usual therapy group, SpO2 was maintained at or above 91% without upper limits. More than two-thirds of the patients included were admitted after surgery or with acute brain disease, but most had a PaO2/FiO2 < 300. In contrast to previous studies, there was no difference in the primary outcome of ventilator-free days or key secondary outcomes including 90- and 180-day mortality, time to discharge, vasopressor-free days, or patients requiring tracheostomy.11
In the same issue, the LOCO2 investigation also was published.12 This trial enrolled 205 patients with ARDS randomized to liberal or conservative oxygen delivery. In this trial, liberal oxygen delivery targeted a PaO2 90 mmHg to 105 mmHg while the conservative target was 55 mmHg to 70 mmHg. The trial was stopped early because of futility and a concern for mesenteric ischemia in the conservative oxygen group. As with the ICU-ROX study, no difference was noted in 28-day mortality; however, the 90-day mortality was 14 percentage points higher in the conservative oxygen group compared with liberal oxygen delivery (44.4% vs. 30.4%; HR, 1.62; 95% CI, 1.02-2.56).12 Five patients died of mesenteric ischemic in the conservative group and none died in the liberal group.
In 2021, the HOT-ICU investigators published a multicenter, randomized study of 2,928 patients with acute hypoxemic respiratory failure.13 The lower oxygenation group targeted a PaO2 of 60 mmHg and the higher oxygenation group targeted PaO2 of 90 mmHg. This trial was powered to have a 90% chance to detect a 5% difference in 90-day mortality. There was no difference in 90-day mortality between the lower oxygen vs. higher oxygen groups (42.9% vs. 42.4%, respectively).13 Furthermore, there was no difference in days alive after hospital discharge, days alive without ventilator support, or serious adverse events. Specifically, there was no difference in intestinal ischemia between the lower and higher oxygen groups (32 vs. 29 patients, respectively).13 Outcomes, including survival and health-related quality of life, were not different in the two groups even at one year of follow-up.14
The publication of these newer trials led to an updated meta-analysis of studies examining higher vs. lower oxygenation strategies.15 These authors included 36 trials totaling 20,166 patients to test the hypothesis that higher oxygenation strategies were associated with increased mortality and adverse events. Of the trials reporting mortality, there was no significant difference between higher and lower oxygenation strategies overall (10.0% vs. 9.7%).15 Within only those studies with low risk of bias, there was no increased risk of mortality with higher vs. lower oxygenation strategies (relative risk [RR], 0.98; 95% CI, 0.89-1.09).15 Furthermore, no difference was noted in serious adverse events when examining all relevant trials or those specifically with low risk of bias. There were no differences in quality of life, lung injury, sepsis, or cardiovascular events, although overall the quality of evidence was low.
Lessons Learned
Recent publications contribute to a growing appreciation that oxygen, like many other interventions, should be prescribed thoughtfully. A summary of these studies suggests several lessons learned. (See Table 1.) First, excessive oxygen delivery does have the potential to cause harm. While some of the earliest observations about oxygen toxicity were confounded by the underlying disease state that prompted the need for supplemental oxygen, recent studies with the highest oxygen targets show the greatest harm. The Oxygen-ICU study allowed PaO2 to rise to 150 mmHg before reducing delivery.3 Similarly, the HYPERS2S study administered 100% oxygen to mechanically ventilated patients regardless of need.6 Re-evaluation of ARDSNet data showed a relationship between excess oxygen delivery and mortality despite multivariate correction for disease severity.5 Those reviews and meta-analyses that include patients admitted with conditions other than respiratory failure who receive supplemental oxygen do not show benefit, and some demonstrate harm.2,4,15 Thus, preventing unnecessary oxygen delivery seems reasonable.
Second, modest targets for oxygenation appear safe. In those studies with oxygen targets ranging from PaO2 55 mmHg to 105 mmHg or SpO2 88% to 97%, no significant differences were noted in outcomes with higher vs. lower oxygenation targets. (See Table 1.)
The ICU-ROX study showed no difference with conservative or usual oxygen delivery on its primary outcome of ventilator-free days or secondary mortality outcomes.11 The LOCO2 trial was stopped early for concern of harm with conservative oxygen delivery, although confidence intervals for 28-day, 90-day, and ICU mortality between the two groups substantially overlap.12 The HOT-ICU trial showed no difference in outcomes at 90-days or with one-year follow-up.13,14 Even the Oxygen-ICU trial that indicated harm with excess oxygenation noted a U-shaped mortality curve suggesting a range of safe O2 targets up to 107 mmHg.3 Harm from oxygen toxicity seemed to occur with the highest SpO2 for the longest periods of time, but this was mitigated by positive end-expiratory pressure (PEEP) commonly used in mechanically ventilated patients.9
The care of critically ill patients in the ICU remains challenging. Recent studies support current protocols for the delivery of oxygen to patients who require supplemental oxygen due to respiratory failure of varying etiologies. Broad oxygenation targets of SpO2 between 88% to 97% continue to be supported by recent trials.
REFERENCES
1. Ashbaugh DB, Bigelow DB, Petty TL, et al. Acute respiratory distress in adults. Lancet 1967; Aug 12:31923.
2. Shah T. Liberal oxygen therapy in the ICU: Time to change practice? Critical Care Alert 2019;26:81-85.
3. Girardis MB, Busani S, Damiani E, et al. Effect of conservative vs conventional oxygen therapy on mortality among patients in an intensive care unit: The Oxygen-ICU randomized clinical trial. JAMA 2016;316:1583-1589.
4. Chu DK, Kim LHY, Young PJ, 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.
5. Aggarwal NB, Brower RG, Hager DN, et al. Oxygen exposure resulting in arterial oxygen tensions above the protocol goal was associated with worse clinical outcomes in acute respiratory distress syndrome. Crit Care Med 2018;46:517-524.
6. Asfar P, Schortgen F, Boisramé-Helms J, et al; The HYPERS2S Investigators. 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.
7. Panwar R, Hardie M, Bellomo R, et al; CLOSE Study Investigators; ANZICS Clinical Trials Group. Conservative versus liberal oxygenation targets for mechanically ventilated patients: A pilot multicenter randomized controlled trial. Am J Respir Crit Care Med 2016;193:43-51.
8. Aggarwal NR, Brower RG. Targeting normoxemia in acute respiratory distress syndrome may cause worse short-term outcomes because of oxygen toxicity. Ann Am Thorac Soc 2014;11:1149-1453.
9. Tretter VZ, Zach ML, Böhme S, et al. Investigating disturbances of oxygen homeostasis: From cellular mechanisms to the clinical practice. Front Physiol 2020;11:947.
10. Griffith DE, Garcia JG, James HL, et al. Hyperoxic exposure in humans: Effects of 50 percent oxygen on alveolar macrophage leukotriene B4 synthesis. Chest 1992;101:392-397.
11. ICU-ROX Investigators and the Australian and New Zealand Intensive Care Society Clinical Trials Group; Mackle D, Bellomo R, Bailey M, et al. Conservative oxygen therapy during mechanical ventilation in the ICU. N Engl J Med 2020;382:989-998.
12. Barrot L, Asfar P, Mauny F, et al; LOCO2 Investigators and REVA Research Network. Liberal or conservative oxygen therapy for acute respiratory distress syndrome. N Engl J Med 2020;382:999-1008.
13. Schjørring OL, Klitgaard TL, Perner A, et al; HOT-ICU Investigators. Lower or higher oxygenation targets for acute hypoxemic respiratory failure. N Engl J Med 2021;384:1301-1311.
14. Crescioli E, Klitgaard TL, Poulsen LM, et al. Long-term mortality and health-related quality of life of lower versus higher oxygenation targets in ICU patients with severe hypoxaemia. Intensive Care Med 2022;48:714-722.
15. Barbateskovic MS, Schjørring OL, Krauss SR, et al. Higher vs lower oxygenation strategies in acutely ill adults: A systematic review with meta-analysis and trial sequential analysis. Chest 2021;159:154-173.
The care of critically ill patients in the ICU remains challenging. Recent studies support current protocols for the delivery of oxygen to patients who require supplemental oxygen due to respiratory failure of varying etiologies. Broad oxygenation targets of SpO2 between 88% to 97% continue to be supported by recent trials.
You have reached your article limit for the month. Subscribe now to access this article plus other member-only content.
- Award-winning Medical Content
- Latest Advances & Development in Medicine
- Unbiased Content