Should O2 exposure be reduced in critically ill patients on MV, O2?

Findings from the UK-ROX* trial do not support reducing oxygen (O₂) exposure by targeting a peripheral O₂ saturation (SpO₂) of 90 percent in adult patients receiving mechanical ventilation (MV) and supplemental O₂ in the intensive care unit (ICU).

In critically ill patients, supplemental O₂ is typically administered via a liberal approach; however, excessive administration may be harmful. [Crit Care 2008;doi:10.1186/cc7150; Am J Respir Crit Care Med 2019;200:1373-1380] “Achieving a balance between too little and too much O₂ [may] be essential to optimize clinical outcomes for patients,” the investigators said.

“[H]owever, there is insufficient evidence to guide its therapeutic use and minimize the potential harm caused by administering too little or too much,” they noted. “[In our study,] minimizing O₂ exposure through conservative O₂ therapy did not significantly reduce all-cause mortality at 90 days … in mechanically ventilated adults receiving supplemental O₂ in ICUs.”

Although the total exposure to supplemental O₂ was 29.3 percent lower in the conservative vs usual O₂ therapy arm (20.3 vs 28.7 [100 percent-equivalent hours, respectively]; difference, -8.4 hours), both groups had similar proportions of participants who had died by day 90 (35.4 percent vs 34.9 percent).

After adjusting for prespecified baseline variables, the risk difference was 0.7 percentage points (p=0.28). The unadjusted risk difference was 0.5 percentage points. [JAMA 2025;334:398-408]

There were also no significant differences between the conservative and usual O₂ therapy arms in terms of secondary mortality outcomes, such as time to death (adjusted hazard ratio [HR], 1.01), duration of acute hospital stay among survivors (median 20 vs 21 days; subdistribution HR, 0.98; p=0.27), and days alive and free from organ support at 30 days (median 16 days for both; proportional odds ratio, 1.01; p=0.64).

The incidences of serious adverse events were also similar between the conservative and usual O₂ therapy arms (0.7 percent vs 0.4 percent).

A total of 16,434 participants (median age 60 years, 38.2 percent women) were included in the primary analysis. Prior to randomization, the median SpO₂ was 97 percent in the conservative O₂ therapy arm and 96 percent in the usual O₂ therapy arm. The participants were randomized after first receiving invasive MV in the ICU (median time to randomization 5 hours).

Participants in the conservative O₂ therapy arm received the lowest fraction of inspired oxygen (FIO₂) possible to maintain their pulse oximeter-derived SpO₂ at 90 percent, while those in the usual O₂ therapy arm received supplemental O₂ at the discretion of the treating physician.

The most common reason for admission was sepsis (33.1 percent), followed by hypoxic-ischaemic encephalopathy (9.2 percent) and acute brain injury (2.2 percent). The rest (55.5 percent) were not in any of the prespecified subgroups of interest. About 7 percent of participants had confirmed or highly suspected COVID-19 on enrolment.

“Our findings add to the understanding of oxygenation targets for critically ill patients by evaluating O₂ therapy in more participants than all prior trials combined,” said the researchers.

The findings from an ongoing randomized trial with a larger cohort are anticipated to shed light on the difference between conservative and protocolized liberal O₂ therapy (minimum acceptable FIO₂ of 0.3). [J Intensive Care Soc 2023;24:399-408]