DESIGN: Post hoc analysis of a pragmatic, open-label, multicenter randomized controlled trial.

SETTING: Fifteen PICUs across the United Kingdom and Scotland.

PATIENTS: Children aged 38 weeks corrected gestational age to 15 years accepted to a participating PICU as an unplanned admission and receiving invasive mechanical ventilation with supplemental oxygen for abnormal gas exchange.

METHODS: Hierarchical regression models for Spo2 and Fio2, and ordinal models for the primary trial outcome of a composite of the duration of organ support at 30 days and death, were used to examine the effects of ethnicity, accounting for baseline Spo2, Fio2, and mean airway pressure and trial allocation.

MEASUREMENTS AND MAIN RESULTS: Ethnicity data were available for 1577 of 1986 eligible children, 1408 (89.3%) of which were White, Asian, or Black. Spo2 and Fio2 distributions did not vary according to Black or Asian ethnicity compared with White children. The trial primary outcome measure also did not vary significantly with ethnicity. The point estimate for the treatment effect of conservative oxygenation in Black children was 0.64 (95% CI, 0.33-1.25) compared with 0.84 (0.68-1.04) in the overall trial population.

METHODS: Oxy-PICU was a pragmatic, multicentre, open-label, randomised controlled trial in 15 UK paediatric intensive care units (PICUs). Children admitted as an emergency, who were older than 38 weeks corrected gestational age and younger than 16 years receiving invasive ventilation and supplemental oxygen were randomly allocated in a 1:1 ratio via a concealed, central, web-based randomisation system to conservative peripheral oxygen saturations ([SpO2] 88-92%) or liberal (SpO2 >94%) targets. The primary outcome was the duration of organ support at 30 days following random allocation, a rank-based endpoint with death either on or before day 30 as the worst outcome (a score equating to 31 days of organ support), with survivors assigned a score between 1 and 30 depending on the number of calendar days of organ support received. The primary effect estimate was the probabilistic index, a value greater than 0·5 indicating more than 50% probability that conservative oxygenation is superior to liberal oxygenation for a randomly selected patient. All participants in whom consent was available were included in the intention-to-treat analysis. The completed study was registered with the ISRCTN registry (ISRCTN92103439).

FINDINGS: Between Sept 1, 2020, and May 15, 2022, 2040 children were randomly allocated to conservative or liberal oxygenation groups. Consent was available for 1872 (92%) of 2040 children. The conservative oxygenation group comprised 939 children (528 [57%] of 927 were female and 399 [43%] of 927 were male) and the liberal oxygenation group included 933 children (511 [56%] of 920 were female and 409 [45%] of 920 were male). Duration of organ support or death in the first 30 days was significantly lower in the conservative oxygenation group (probabilistic index 0·53, 95% CI 0·50-0·55; p=0·04 Wilcoxon rank-sum test, adjusted odds ratio 0·84 [95% CI 0·72-0·99]). Prespecified adverse events were reported in 24 (3%) of 939 patients in the conservative oxygenation group and 36 (4%) of 933 patients in the liberal oxygenation group.

INTERPRETATION: Among invasively ventilated children who were admitted as an emergency to a PICU receiving supplemental oxygen, a conservative oxygenation target resulted in a small, but significant, greater probability of a better outcome in terms of duration of organ support at 30 days or death when compared with a liberal oxygenation target. Widespread adoption of a conservative oxygenation saturation target (SpO2 88-92%) could help improve outcomes and reduce costs for the sickest children admitted to PICUs.