Oxygen therapy in the critically ill: Less is the new more?

The evolution of medicine is not limited to the development of new therapies albeit extends to the most pragmatic application of those existing. As the concept of precision medicine dawns, the fraternity is equally focusing at the harms of overzealous treatment. The paradigm shift of “less is the new more” is already making a mark in the modern medicine along the lines of a great Paracelsus saying: “All things are poisons, for there is nothing without poisonous qualities. It is only the dose which makes a thing poison.”

Appropriate to the aforementioned context, oxygen, the most ubiquitously used therapeutic agent, essentially classifies as a drug with characteristic physiological–biochemical properties and a range of effective dosages. Oxygen therapy (administered with the fundamental aim of minimizing cellular hypoxia) can potentially demonstrate peculiar deleterious effects at higher concentrations. Although the physicians administer oxygen therapy in diverse clinical scenarios, such as resuscitation, perioperative period, and intensive care unit (ICU), the considerations for the detrimental effects of prolonged hyperoxia in the critically ill cohort become manifold.

There is a substantial recent literature accumulating in this research area. The oxygen-ICU randomized clinical trial (RCT) involving 480 critically ill participants (with an expected ICU stay ≥72 h) outlined a significantly lower mortality rate of 11.6% with a conservative oxygenation approach [partial pressure of arterial oxygen (PaO₂) of 70–100 mmHg or 94–98% target arterial oxygen saturation (SpO₂)] compared to a 20.2% mortality rate in the more liberal regime allowing a PaO₂ upto 150 mmHg or 97–100% target SpO₂.[1] In addition, a broadly inclusive systematic review and meta-analysis, “Improving Oxygen Therapy in Acute illness” comprising of 25 RCTs amounting to a total of 16,037 patients with underlying critical illness (sepsis, myocardial infarction, stroke, post-cardiac arrest, emergency surgery, etc.) demonstrated a high-quality evidence for the mortality reduction effect of conservative oxygenation robust to a subsequent trial sequential analysis.[2]

The clinical applicability of the meta-analysis results has been interrogated by the practitioners citing a wide range of heterogeneity of the included RCTs with a few RCTs employing a much more liberal oxygenation approach than the usual care. Withstanding this fact, the “Intensive Care Unit Randomized Trial Comparing Two Approaches to Oxygen Therapy” (ICU-ROX) revealed an insignificant difference in 90-day mortality in a comparative evaluation of the conservative oxygenation strategy (91–97% SpO₂) with a usual-care strategy (91–100% SpO₂). Despite comparable mortality between the two strategies, considerable treatment-effect heterogeneity was appreciated wherein the hypoxic–ischemic encephalopathy subset demonstrated favorable outcomes in the background of conservative oxygenation.[3] This positive modulatory impact has been previously depicted in a multicenter cohort study outlining a heightened mortality with post-resuscitation hyperoxia which is often attributed to the accentuation of secondary injury owing to an enhanced oxidative stress.[4] Moreover, the usual care in this trial did not classify as a hyperoxemic or a liberal oxygen strategy which is in sharp contrast to the previous investigations such as the oxygen-ICU trial.

Prompted by the potential bactericidal properties of oxygen therapy, 442 septic subjects were exposed to a 1.00 fraction of inspired oxygen (FiO₂) for the 1^st 24 h in the “Hyperoxia and Hypertonic Saline in Patients with Septic Shock”(HYPERS2S) trial.[5] It is noteworthy that the trial had to be prematurely terminated in view of the unacceptably high mortality rate in the hyperoxia intervention group. However, a post-hoc analysis of 251 septic patients from the ICU-ROX trial suggested a 7% higher mortality in the conservative oxygenation group supporting the notion of a beneficial impact attributable to higher oxygen-thresholds in sepsis.[6] Interestingly, their analysis was not powered enough to detect the suggested effects precluding the clinical extrapolation of the aforementioned findings. In addition, an updated meta-analysis including 17 RCTs failed to demonstrate a strong evidence of reduction in the surgical site infections with a higher perioperative FiO₂ (0.80) when compared to a lower FiO₂ (0.30-0.35).[7]

To conclude, the presumption of hyperoxia as a panacea is being increasingly challenged. While the most recent Cochrane review in this subject suggests that the elevated fractions of oxygen supplementation may incur an accentuated ICU mortality risk,[8] many debate the generalization potential of these results given the heterogeneous settings of investigation and an inherent research reliance on subgroup analysis, often rendering it difficult to distinguish “signals” from “noise” in this peculiarly predisposed cohort.[9] Nevertheless, the acknowledgement to the precision concept in oxygen therapy[10,11] and an improved characterization of the therapy targets is definitely the need of the hour for optimizing the outcomes in the critically ill.