The downside of oxygen therapy examined in critically ill children

Through a range of mechanisms, delivery of oxygen to tissues can be compromised in most conditions that warrant admission to critical care, so that ensuring adequate tissue oxygenation – with the help of supplemental oxygen when necessary – is a prime objective of acute/critical care. Of all therapies used in the management of critically ill patients, supplemental oxygen is often cited as the most common. The objective of oxygen therapy is correction of hypoxemia (reduced blood oxygen) and avoidance of tissue hypoxia (reduced oxygen in tissues).

Although supplemental oxygen has life-preserving benefit for critically/acutely ill patients with hypoxemia and those at risk of hypoxia due to reduced cardiac output, it is associated with risk of oxygen toxicity consequent on hyperoxemia (raised blood oxygen).

Blood gas analysis includes measurement of partial pressure of oxygen in arterial blood (pO₂(a)) and thereby provides the means for helping to identify the need for supplemental oxygen therapy and monitoring its use. In health, breathing room air that results in a fraction of inspired oxygen (FO₂(I)) of 0.21, pO₂(a) is maintained within the normoxemic range 10.6-13.3 kPa (80-100 mmHg). Hypoxemia can thus be defined as pO₂(a) <10 kPa (<80 mmHg) and hyperoxemia can be defined as pO₂(a) >13.3 kPa (>100 mmHg).

Administration of supplemental oxygen therapy necessarily increases FO₂(I) beyond 0.21 (may be greater than 0.60 depending on the prescription) so that hyperoxemia (defined as pO₂(a) >13.3 kPa) is an inevitable consequence of oxygen therapy for many patients.

Over the past decade or so there has been an increasing research interest in establishing whether hyperoxemia induced by supplemental oxygen therapy poses risk for patients. There is now evidence to suggest that severe hyperoxemia (usually defined as pO₂(a) >40 kPa or >300 mmHg) has deleterious effect in terms of increased morbidity and mortality. Most of this evidence is derived from study of adult patients.

This highlighted recently published study sought to test the hypothesis that severe hypoxemia (defined as pO₂(a) >40 kPa (>300 mmHg) is associated with increased mortality risk for critically ill children.

The study was conducted at a quaternary care pediatric intensive care unit (PICU) in Pittsburgh, US. Investigators evaluated all encounters in children (mean age 7.5 years – range 1 month-18 years) admitted to the unit over a 10-year period (2008-2018) and extracted the clinical details of all 6250 children who had at least one documented pO₂(a) value.

Of the 6250 children, 4559 (72.9 %) did not have any instance of severe hyperoxemia and 1691 (27.1 %) had a least one instance of severe hypoxemia, of whom 236 had two instances, and 201 had three or more instances, each separated by an interval of more than 3 hours.

Illness severity was assessed for each of the 6250 patients using the well-validated and internally calibrated Modified Pediatric Logistic Organ Dysfunction-2 (mPELOD-2) score based on the results of a range of routine laboratory tests and clinical measurements. mPELOD-2 has a score range of 0 (no organ dysfunction) to 31 (greatest amount of organ dysfunction).

After adjusting for illness severity and other covariates, severe hyperoxemia was shown to be independently associated with in-hospital mortality (adjusted Odds Ratio 1.78). Adjusted Odds Ratio (aOR) was found to increase with the number of episodes of severe hyperoxemia (aOR – 1.47 for one episode, aOR – 2.01 for two episodes and aOR – 2.53 for three or more episodes).

In separate analysis of recovered pO₂(a) values, it was shown that the higher the pO₂(a), the greater was the chance that observed mortality would be higher than that estimated from the patient’s condition. So that, for example, 158 patients had extreme hyperoxemia (i.e. pO₂(a) equal to or greater than 73.3 kPa, 550 mmHg). The m-PELOD-2 score for each of these patients indicated that the estimated proportional mortality for this group of patients was 0.18, whereas in fact the observed (actual) proportional mortality was 0.3. By contrast, observed mortality broadly reflected expected mortality for patients without severe hyperoxemia (i.e. pO₂(a) <40kPa, 300 mmHg).

This large study involving over 6000 patients, analysis of over 100,000 pO₂(a) results, and application of a range of statistical tools provides robust evidence for the authors to conclude that severe hyperoxemia is independently associated with in-hospital mortality in critically ill children and adolescents. In discussion of their study, the authors reflect on how results of their study fit in to existing evidence of association between severe hyperoxemia and increased mortality/morbidity in adults. In so doing, they provide a broad overview of the topic, including potential impact for revision of guidelines on the use of supplemental oxygen in critical care.