Is increased blood oxygen (pO₂(a)) harmful – the potential adverse effect of oxygen therapy reviewed

One of the utilities of arterial blood gas analysis (BGA) is to assess patient blood oxygenation status. The two principal parameters generated during BGA that allow this assessment are: partial pressure of oxygen in arterial blood (pO₂(a)) and the percentage of functional hemoglobin in arterial blood that is saturated with oxygen (sO₂(a)). In health, when breathing ambient air (i.e. fraction inspired oxygen, FO₂(I) ∼21 %), pO₂(a) is maintained within the approximate reference interval of 10.6-13.3 kPa (80-100 mmHg); the equivalent reference interval for sO₂(a) is 94-99 %. 

Reduced pO₂(a)/sO₂(a), which implies impaired gas exchange in the lungs due to reduced alveolar ventilation and/or perfusion, is a feature of a range of acute and chronic respiratory conditions, including acute asthma, pneumonia and chronic obstructive pulmonary disease. The associated hypoxemia (reduced oxygen in blood) threatens adequate delivery of oxygen to tissue cells, with consequent risk of damaging (potentially lethal) global tissue hypoxia. 

By contrast, increased pO₂(a) in association with maximally saturated hemoglobin (i.e. sO₂(a) 100%), a condition called hyperoxemia, has no pathological cause, and only arises iatrogenically as a result of supplemental oxygen therapy administered to all patients at risk of tissue hypoxia. Whilst this patient group includes those with significant reduction in pO₂(a)/sO₂(a), it is not restricted to these patients. Adequate oxygenation of tissues depends not only on adequate blood oxygenation but also on adequate tissue perfusion, i.e. normal cardiac output. Patients with circulatory deficiency (due to e.g. heart failure) are at risk of tissue hypoxia despite adequate blood oxygenation and normal pO₂(a)/sO₂(a).

Administration of supplemental oxygen to patients with normal respiratory function, whose pO₂(a)/sO₂(a) is within normal limits, inevitably results in supranormal levels (hyperoxemia) because FO₂(I) is greater than 21 %; depending on the mode and rate of delivery, oxygen therapy can result in FO₂(I) that ranges from 25-100 %. 

Early administration of supplemental oxygen has been a routine of emergency/ critical care for most patients. This is not necessarily guided by knowledge of patient’s blood oxygenation status (pO₂(a)/sO₂(a)) and consequently can result in hyperoxemia (raised pO₂(a)). Prevention of tissue hypoxia is the prime motivation of supplemental oxygen therapy, and hyperoxemia has historically been viewed as a harmless, perhaps beneficial temporary side effect that should be tolerated to prevent the greater harm, tissue hypoxia. 

Accumulating evidence in recent years suggests that hyperoxemia during critical illness is not as benign as once supposed. In a recently published article, three Italian intensivists with research interest in the potential harmful effects of oxygen therapy, review this evidence. 

The authors begin the article by making the distinction between hyperoxia, increase in oxygen supply (FO₂(I) >21 %) and hyperoxemia, increased oxygen in blood (pO₂(a) >100 mmHg (13.3 kPa). They report that in observational studies, various cut-offs have been used to define hyperoxemia, ranging from 120 mmHg (16 kPa) to as high as 300 mmHg (40 kPa).

The relationship between hyperoxemia and patient outcome has been investigated in a number of studies that have revealed a linear relationship between increasing oxygen tensions and mortality. A recent 2017 study found that although severe hyperoxemia (defined as pO₂(a) >200 mmHg) was more consistently associated with adverse outcome, mortality was found to increase linearly with length of time patients were exposed to a milder degree of hypoxemia (pO₂(a) in the range of 120-200 mmHg).

A discussion of the biological effects of hyperoxemia follows. The main focus here is animal and cell line studies of excess oxygen exposure. These have demonstrated, for example, that the lungs of mice exposed to excess oxygen suffer dose-dependent, time-dependent pro-inflammatory tissue damage. Detail of the mechanism of the oxidative lung damage revealed by these studies is discussed. 

Some animal studies have revealed that excessive oxygen can have adverse systemic effect beyond the lungs. In an animal model of sepsis, for example, 24-hour exposure to hyperoxemia was associated with increased serum concentration of damaging reactive oxygen species and inflammatory cytokines, greater spread of infection and worse multiple organ dysfunction. Other adverse systemic effects of hyperoxemia identified by animal and/or human study include: vasoconstriction, reduction in coronary blood flow, and reduced microvascular perfusion (with paradoxical reduction on regional oxygen delivery). 

The next section of this article focuses on a number of recent clinical studies performed in an intensive/critical care setting that were all designed to compare outcome of critically ill patients exposed to conservative protocols of oxygen administration aimed at reducing the incidence and severity of hyperoxemia, with outcome of patients exposed to more liberal use of oxygen. In summary, these have revealed that using a relatively low oxygen saturation target (SpO₂ 90-95 %) to guide the use of oxygen therapy – thereby reducing the risk of hyperoxemia – is associated with better outcome by a number of measures, including: reduced risk of non-respiratory organ failure; lower lactate levels; reduction in the time patients need to be mechanically ventilated; and reduced hospital mortality. 

The final sections of this article focus on studies examining the efficacy of supplemental oxygen in the treatment of three specific subsets of critically ill patient: those suffering stroke or traumatic brain injury; those suffering myocardial infarction or cardiac arrest; and finally, those with sepsis. 

The notion suggested by early preclinical studies that hyperoxemia induced by administration of high-flow oxygen is beneficial to victims of stroke and brain injury is challenged by the results of more recent study cited in this article that reveal an association between hyperoxemia and mortality among these patients. 

The once routine administration of supplemental oxygen to all patients suffering myocardial infarction (MI) is no longer recommended because of recent evidence that hyperoxia/hyperoxemia has the effect of reducing coronary blood flow and causing a paradoxical increase in myocardial ischemia that exacerbates ischemic injury induced during myocardial infarction. Current guidelines now advise that oxygen therapy should only be routinely administered to MI patients if there is clear evidence of hypoxemia (sO₂(a) <90 %). 

The authors of this wide-ranging literature review conclude by cautioning that supplemental oxygen delivered to non-hypoxemic patients leads to only marginal increase in delivery of oxygen to tissues but at the expense of a range of hyperoxemic-mediated adverse effects. They find no evidence to support the liberal use of supplemental oxygen therapy to any subset of critically ill patients. It is important to avoid hyperoxemia by reducing exposure of patients to unnecessary oxygen supplementation. To this end they recommend that oxygen therapy should be precisely titrated against patient-appropriate pO₂(a)/sO₂(a) targets.