top of page

BOX Trial (Oxygenation)

BOX Trial (Oxygenation)


August 27, 2022

Oxygen Targets in Comatose Survivors of Cardiac Arrest

Mazen Kherallah

Summarized by: 

What was the research question?

  • In patients who remain comatose after out-of-hospital cardiac arrest, does a restrictive oxygenation target improve outcomes compared to a liberal oxygenation target?

How did they do it?

  • A double-blind, dual-center, randomized trial with a 2-by-2 factorial design in Denmark.

  • A total of 789 comatose adults who had been resuscitated after an out-of-hospital cardiac arrest of presumed cardiac cause were randomized in a 1:1 ratio to receive a restrictive oxygen target of a partial pressure of arterial oxygen (Pao2) of 9 to 10 kPa (68 to 75 mm Hg) (394 patients) or a liberal oxygen target of a Pao2 of 13 to 14 kPa (98 to 105 mm Hg) (395 patients).

  • Patients were also assigned to one of two blood pressure targets and results were reported separately.

  • The primary outcome was a composite of death from any cause or hospital discharge with a severe disability or coma within 90 days.

  • Secondary outcomes included neuron-specific enolase levels at 48 hours, death from any cause, scores on the Montreal Cognitive Assessment, modified Rankin scale at 90 days, and the CPC at 90 days.

What did they find?

  • The primary-outcome was not significantly different in the restrictive-target group compared to the liberal-target group (32% vs 33.9%, hazard ratio, 0.95; 95% confidence interval, 0.75 to 1.21; P=0.69).

  • At 90 days, mortality was not different between the two groups (28.7% vs 31.1%).

  • In both groups, the median CPC was 1; the corresponding median modified Rankin scale scores were 2 in the restrictive-target group and 1 in the liberal-target group, and the corresponding median Montreal Cognitive Assessment scores were 27 in both groups. The median neuron-specific enolase level at 48 hours was 17 μg per liter in the restrictive-target group and 18 μg per liter in the liberal-target group.

  • Adverse events were not significantly different between the groups.

Are there any limitations?

  • Follow-up was a challenge due to COVID restrictions.

  • Generalizability may be affected as the trial was conducted at two centers only and included a population of patients with a high prevalence of acute coronary syndrome and a relatively good prognosis.

  • In some patients, the spontaneous Pao2 value may have exceeded the restrictive oxygenation target.

What does it mean?

  • Restrictive oxygenation target did not reduce mortality rate, severe disability, or coma in comatose patients who had been resuscitated after out-of-hospital cardiac arrest.


bottom of page