Abstract & Commentary

Is Pulmonary Oxygen Toxicity Still a Clinically Relevant Issue?

By Richard H. Kallet, MS, RRT, FAARC, FCCM, Director of Quality Assurance, Respiratory Care Services, San Francisco General Hospital, is Associate Editor for Critical Care Alert.

Dr. Kallet reports no financial relationships relevant to this field of study.

Synopsis: This retrospective study examined pulse oximetry saturation and inspired oxygen fraction at 15-minute intervals in patients with acute lung injury. Using predetermined cutoff values for each variable, the authors determined that the duration of excessive oxygen administration was significantly related to worse oxygenation index at 48 hours of mechanical ventilation.

Source: Rachmale S, et al. Practice of excessive inspired oxygen supplementation and effect on pulmonary outcomes in mechanically ventilated patients with acute lung injury. Respir Care 2012; May 15. [Epub ahead of print.]

Rachmale and colleagues retrospectively identified 210 patients with acute lung injury/acute respiratory distress syndrome (ALI/ARDS) undergoing mechanical ventilation for longer than 48 hours. They abstracted matched measurements of pulse oximetry saturation (SpO2) and inspired oxygen fraction (FiO2) at 15-minute intervals. Excessive O2 exposure was defined a priori as an FiO2 greater than 0.5 when the corresponding SpO2 was greater than 92%. By institutional guidelines, tidal volumes ostensibly were kept between 6-8 mL/kg predicted body weight, whereas oxygenation goals were set by clinician discretion. Initially, both groups had comparable tidal volumes and positive end expiratory pressure (PEEP) levels, similar hemoglobin levels, and hemodynamic status.

In the first 48 hours of mechanical ventilation, 74% of patients had a SpO2 greater than 92% with a corresponding FiO2 greater than 0.5, whereas 53% had exceeded the SpO2 threshold on an FiO2 greater than 0.70. The mean duration of excessive oxygen exposure was 17 hours (interquartile range, 7.5-33 hours). Compared to the cohort of patients defined as having appropriate FiO2 exposure, those with an excessive FiO2 exposure also had a significantly higher median oxygenation index (OI), which calculated as (O2% × mean airway pressure/arterial oxygen tension [5.1 vs 13.3, respectively; P = 0.0001]), and also had a significantly greater change in OI from baseline to 48 hours (-1.5 vs 4.6, respectively; P = 0.0001).

It was determined that at least 12 hours of excessive FiO2 exposure was required to produce a significant change in OI, after which further exposure had a linear relationship with OI found at 48 hours. The correlation between excessive FiO2 and deterioration in OI remained despite adjusting for baseline OI and illness severity, as well as the degree of excessive exposure (i.e., FiO2 > 0.50 or 0.70). Overall, 74% of patients had an excessive O2 exposure representing approximately 20 hours of the first 48 hours of mechanical ventilation. Patients exposed to excessive O2 had longer durations of mechanical ventilation, ICU, and hospital length of stay. Mortality was not different.


The severity of hyperoxic acute lung injury (HALI) is directly proportional to alveolar O2 partial pressure (PAO2) and exposure duration. Across numerous species with normal lungs, inflammatory changes typically occur in a dose-dependent fashion only when PAO2 exceeds 450 mmHg (FiO2 of 0.60) for several days.1 Recently, it was found that mitochondrial production of reactive O2 species increases at a substantially greater rate when FiO2 is greater than 0.60.2 HALI is both species-dependent and varies widely among individual animals. Susceptibility reflects a genetically determined response to injury through a balance between complex pro- and anti-inflammatory cellular mechanisms.1 Humans appear to be relatively more resistant than other mammals to HALI.

From the late 1960s through the 1970s, the risk of HALI was exaggerated, in part because the pathogenesis of ARDS was poorly understood and its connection to ventilator-induced lung injury was unknown. In addition, all evidence used to support the clinical incidence of HALI was based on small, poorly controlled, retrospective studies. Moreover, as PEEP became a standard therapy for improving oxygenation in ARDS, most patients could be managed with a relatively non-toxic FiO2, and consequently concern over HALI faded as a clinical priority.

However, whether excessive FiO2 exposure might potentiate ALI/ARDS remains unknown. The few animal studies examining this issue have produced mixed results; likely because different animals and different methods of inducing ALI were used.2 Currently, there is no compelling evidence supporting the findings of Rachmale et al. And as an initial impression, both the level of FiO2 and the duration of excessive O2 exposure in this study seem inadequate to substantially augment ALI/ARDS. The investigators also were appropriately cautious in their interpretation given the retrospective nature of their study, and, most importantly, noting that mathematical coupling necessarily biased their results (as inspired O2% is a component of OI).

However, the statistical significance of their results is intriguing and deserves further study. Their findings would have been more convincing had they reported other measures associated with worsening ALI/ARDS (e.g., dead-space fraction and pulmonary compliance). But the authors' findings raise some important questions. First, should severe ARDS be managed more aggressively with higher PEEP and other lung recruitment strategies that might contribute to improved outcomes by reducing HALI? Recent meta-analyses of prospective randomized, controlled trials suggest that higher PEEP3 and prone positioning4 may confer a survival benefit in the subset of patients with very severe ARDS. Second, as advances continue to be made in closed-loop mechanical ventilation, should automated control of FiO2 be used to minimize periods of excessive O2 exposure? As Rachmale et al have pointed out, clinicians are reluctant to maintain unstable patients at the brink of desaturation and prefer to maintain a "buffer zone" in O2 saturation. This is particularly true when the acuity and census are high so that the response time to correct episodes of hypoxemia might be delayed.

Rachmale and colleagues have refocused attention on the possibility that HALI may contribute to morbidity in ALI/ARDS. It is possible that today the pendulum may have swung too far away from the excessive concern over HALI that existed in the 1970s. The critical care community at large has not enthusiastically embraced the routine use of high PEEP and other ancillary therapies that have been shown to improve oxygenation in ARDS. However, in severe ARDS, prolonged exposure to a FiO2 of 0.80 or greater produces unambiguously toxic effects that may complicate patient management and prolong the need for mechanical ventilation. And in the subset of patients with severe ARDS in whom HALI is most likely to have a profound negative impact, we will likely never have convincing proof. For example, estimates are that it would take 20 randomizing centers approximately 15 years to provide confirmation of the recent meta-analysis that prone positioning improves mortality in very severe ARDS.4 Therefore, the debate over how to balance PEEP and FiO2 will continue, but perhaps in the near future with slightly more regard for the parsimonious use of O2 therapy.


  1. Kallet RH, Matthay MA. Hyperoxic acute lung injury. Respir Care 2013;58(1). [In press.]
  2. Turrens JF. Mitochondrial formation of reactive oxygen species. J Physiol 2003;552:335-344.
  3. Briel M, et al. Higher vs. lower positive end-expiratory pressure in patients with acute lung injury and acute respiratory distress syndrome. Systematic review and meta-analysis. JAMA 2010;303:865-873.
  4. Gattinoni L, et al. Prone positioning improves survival in severe ARDS: A pathophysiologic review and individual patient meta-analysis. Minerva Anestesiol 2010;76:448-454.