In addition to mechanical factors which damage the lung, high inspired concentrations of oxygen are believed to be toxic. Most of the evidence for this comes from animal studies, as pathological samples from patients may be difficult to differentiate from the underlying disease. However, it is possible that animal models of oxygen toxicity do not reflect pathology in humans.
A prospective study of 10 patients with irreversible brain damage (five ventilated with an FiO 2 of 0.21, and the rest at an FiO2 of 1.0) found that after 40 h the group ventilated with pure oxygen had significantly poorer oxygenation compared with the 'air' group (temporarily measured at FiO 2 1.0 for comparison). The intrapulmonary shunt and ratio of dead-space to tidal volume were also significantly increased in the 'oxygen' group.
Acute tracheobronchitis with substernal distress may begin within 4 to 22 h of breathing 100 per cent oxygen in normal subjects; in contrast, others have found that short exposure (24-48 h) to 100 per cent oxygen appeared to do little harm, with no change in intrapulmonary shunt, dead-space, or clinical outcome following cardiac surgery. However, normal subjects may not reflect the situation in patients with diseased lungs. The individual patient's liability to oxygen toxicity may depend on his or her age, nutritional status, and endocrine status, as well as previous exposure to oxygen or other oxidants ( Deneke and Fanburg 1980).
The changes of oxygen toxicity tend to be rather non-specific, i.e. atelectasis, inflammation, edema, alveolar hemorrhage, fibrin deposition, and thickening and hyalinization of alveolar membranes.
High FiO2 may cause atelectasis principally by two mechanisms. Firstly, in partially obstructed airways, oxygen is absorbed from the alveoli faster than it is replenished from the upper airways. This tendency will be exaggerated if FiO 2 is 1.0 and no nitrogen is present to splint the alveoli. Secondly, in the presence of high concentrations of oxygen, surfactant may be deficient in amount and function.
Further toxic effects of oxygen are believed to be mediated through what are loosely termed oxygen radicals, the most important of which are superoxide (O 2-), hydrogen peroxide (H2O2), singlet oxygen (1O2), and hydroxyl radical (OH-). These may cause lipid peroxidation, protein sulfhydryl oxidation, and DNA damage, leading to cell death. The protective mechanisms against these radicals under normal circumstances (e.g. superoxide dismutase, catalase, and reduced glutathione) may be overwhelmed in the diseased state. It is not yet clear whether giving reducing agents such as ^-acetylcysteine or superoxide dismutase has any therapeutic benefit. These toxic effects may lead to an inflammatory response, possibly resulting in edema, fibrin deposition with hyalinization of alveolar membranes, and alveolar hemorrhage.
The 'safe FiO2' in humans is unclear. Empirical animal evidence suggests that the dose of oxygen may be a function of partial pressure and length of exposure. In humans, FiO2 values of 50 to 60 per cent are generally regarded to be the upper limit of 'safety'. However, in view of the lack of unequivocal evidence about the significance of oxygen toxicity (possibly rarely of practical importance), it is perhaps best to adopt a common-sense approach, i.e. fear of causing oxygen toxicity should not take precedence over the relief of arterial hypoxemia in mechanically ventilated patients, while using the minimum FiO 2 to provide an acceptable PaO2 of 8 kPa.
Was this article helpful?