Oxygen saturation in critical illness: could low-normal be best?

Oxygen is essential for life, but by forming superoxides and free radicals, supplemental oxygen can also inflict damage on lung and other body tissues.

The sweet spot for oxygen delivery in critically ill patients is unknown, but increasing evidence suggests that when it comes to blood oxygen saturation during critical illness, "normal" levels might actually be harmful.

High levels of blood oxygen have been associated with lower survival after cardiac arrest; worsened injury from myocardial infarction; and poorer outcomes from critical illness generally.

A provocative randomized trial in JAMA (published as a "preliminary communication" owing to its serious limitations) adds to the swirl of suspicion of harm that might result from high levels of arterial oxygen during critical illness.

Authors tried to randomize 660 critically ill patients at a single ICU in Modena University Hospital, Italy to either low normoxia (SaO2 94-98%) or high normoxia (SpO2 97-100%). Only those expected to stay 3 days, without COPD exacerbations or ARDS were enrolled. All-cause ICU mortality in those who stayed 3 days was the primary outcome.

After 480 patients were enrolled, the study was stopped when an earthquake damaged the hospital and an interim analysis showed a surprisingly large signal: the patients in the low-normal oxygen group had an almost 9% absolute risk reduction for ICU mortality (20.2% vs 11.6%). Many more patients in the low-normal oxygen group survived to hospital discharge, as well.

The groups were imbalanced at baseline, with less sick patients in the low-normal oxygen group probably boosting the observed effect. Further, the study was underpowered due to the low enrollment. Also, a strict intention-to-treat analysis was not followed, because patients who did not stay in ICU for 72 hours were enrolled but excluded from analysis.

For all of these reasons, this is not a conclusive study by any stretch of the imagination. But given the mounting suggestion of possible harm from high normal or supranormal arterial oxygen levels, and the apparent absence of harm from low-normal oxygenation, it seems prudent to target oxygen saturation in the mid-90s percent for many, if not most critically ill patients.

Read more:

Effect of Conservative vs Conventional Oxygen Therapy on Mortality Among Patients in an Intensive Care Unit. The Oxygen-ICU Randomized Clinical Trial. JAMA. 2016 Oct 18;316(15):1583-1589.