Press "Enter" to skip to content

Trial: No Need for Intensive Oxygen in the ICU

Lowering the oxygen-saturation target in intensive care unit (ICU) patients showed no disadvantage relative to usual care either in the number of days patients needed mechanical ventilation or in overall survival, according to a randomized trial of 1,000 patients in Australia and New Zealand.

For the main study outcome of ventilator-free days, the medians were 21.3 and 22.1 days, respectively, in the conservative and usual oxygen-delivery groups for an absolute difference -0.3 days (95% CI -2.1 to 1.6; P=0.80), reported Paul Young, MD, of the Medical Research Institute of New Zealand, who presented the findings at the World Congress of Intensive Care in Melbourne, Australia.

And at 180 days, 35.7% and 34.5% of patients died in the conservative therapy and usual therapy groups, respectively (unadjusted odds ratio 1.05; 95% CI 0.81-1.37; hazard ratio 1.05, 95% CI 0.85-1.30).

The results, also published simultaneously in the New England Journal of Medicine, are at odds with those of a single-center study from Italy, published in JAMA in 2016, which was stopped early after interim analysis suggested an increase in deaths associated with higher-dose oxygen therapy.

The concern has been that high levels of blood oxygen resulting from supplementation may harm patients via systemic and pulmonary oxidative injury, Young and colleagues explained. That has prompted efforts to identify supplemental oxygen regimens that don’t oversaturate while still preventing hypoxemia.

In the trial, “conservative” oxygen delivery involved setting pulse oximetry (Spo2) alarms to keep oxygen saturation from 90% to 97%. When fraction of inspired oxygen (Fio2) decreased to 0.21 (ambient, non-therapeutic oxygen delivery), supplemental oxygen was discontinued in extubated patients as long as Spo2 remained above the 90% lower limit.

Usual care did not require limiting Fio2 or Spo2 other than keeping the latter above 90%, although Fio2 less than 0.3 was “discouraged” during mechanical ventilation. Upper limits on Spo2 were prohibited.

Young said the new trial data provide “a degree of reassurance” that the findings from the Italian trial represented a false-positive result.

He told MedPage Today that, based on the latest findings, “a prudent approach [to oxygen delivery] for patents on mechanical ventilation in the ICU is the middle ground where one should neither systematically target high levels of oxygen nor aggressively reduce oxygen in a way that exposes patients to hypoxemia.”

Subgroup analysis did show that conservative oxygen therapy had a benefit for patients who had suffered sudden cardiac arrest or had other anoxic brain injuries.

Niall Ferguson, MD, of Toronto General Hospital, who was not involved with the study, said the strong signal of better outcomes suggests that conservative oxygen delivery should be the standard of care for this subset of patients. “There is also observational data showing that avoiding high levels of oxygen in the setting of cardiac arrest seems to improve patient outcomes,” he told MedPage Today.

Study details

The study’s 1,000 ICU patients all expected to need mechanical ventilation beyond the day after recruitment.

Assigned oxygen therapy was delivered until discharge from the ICU or until 28 days after randomization.

Other findings included:

  • Among the patients with suspected hypoxic-ischemic encephalopathy, the median number of ventilator-free days was 21.1 in the conservative oxygen group vs none in the usual care group (absolute difference 21.1 days, 95% CI 10.4-28.0)
  • In post-hoc analysis of patients with suspected hypoxic-ischemic encephalopathy performed at 180 days, 43% of patients in the conservative oxygen group died, compared with 59% in the usual oxygen delivery group (RR 0.73, 95% CI 0.54-0.99; HR 0.67, 95% CI 0.43-1.03)

The trial results do not preclude “the possibility of benefit or harm with more liberal oxygen regimens than those used in our usual-oxygen group,” said Young and colleagues.

“Our findings decrease the probability that the use of our protocol for conservative oxygen therapy in this population would result in markedly lower mortality than the use of usual oxygen therapy. However, the confidence intervals around our mortality estimates are sufficiently wide that we cannot rule out important effects of our conservative oxygen regimen on mortality,” the team wrote.

Young said future trials should be powered to detect a true minimally important difference in mortality.

He said the researchers are now working to secure funding for a much larger international trial “designed to test the hypothesis that, compared with usual oxygen therapy, conservative use of oxygen reduces mortality by 1.5 percentage points or greater in adults who are ventilated following an unplanned ICU admission or emergency intubation in the ICU.”

The research was funded by the New Zealand Health Research Council.