Conservative versus liberal oxygenation targets in critically ill children: the randomised multiple-centre pilot Oxy-PICU trial

Mark J. Peters, Gareth A. L. Jones, Daisy Wiley, Jerome Wulff, Padmanabhan Ramnarayan, Samiran Ray, David Inwald, Michael Grocott, Michael Griksaitis, John Pappachan, Lauran O’Neill, Simon Eaton, Paul R. Mouncey, David A. Harrison, Kathryn M. Rowan

Oxygen saturation monitoring for children receiving respiratory support is standard worldwide. No randomised clinical trials have compared peripheral oxygen saturation (SpO2) targets for critically ill children. The harm of interventions to raise SpO2 to > 94% may exceed their benefits. We undertook an open, parallel-group randomised trial of children > 38 weeks completed gestation and < 16 years of age receiving invasive or non-invasive respiratory support and supplemental oxygen who were admitted urgently to one of three paediatric intensive care units. A 'research without prior consent' approach was employed. Children were randomly assigned to a liberal oxygenation group (SpO2 targets > 94%) or a conservative oxygenation group (SpO2 = 88-92% inclusive). Outcomes were measures of feasibility: recruitment rate, protocol adherence and acceptability, between-group separation of SpO2 and safety. The Oxy-PICU trial was registered before recruitment: ClinicalTrials.gov identifier NCT03040570. A total of 159 children met the inclusion criteria, of whom 119 (75%) were randomised between April and July 2017, representing a rate of 10 patients per month per site. The mean time to randomisation from first contact with an intensive care team was 1.9 (SD 2.2) h. Consent to continue in the study was obtained in 107 cases (90%); the children's parents/legal representatives were supportive of the consent process. The median (interquartile range, IQR) of time-weighted individual mean SpO2 was 94.9% (92.6-97.1) in the conservative oxygenation group and 97.5% (96.2-98.4) in the liberal group [difference 2.7%, 95% confidence interval (95% CI) 1.3-4.0%, p < 0.001]. Median (IQR) time-weighted individual mean FiO2 was 0.28 (0.24-0.37) in the conservative group and 0.37 (0.30-0.42) in the liberal group (difference 0.08, 95% CI 0.03-0.13, p < 0.001). There were no significant between-group differences in length of stay, duration of organ support or mortality. Two prespecified serious adverse events (cardiac arrests) occurred, both in the liberal oxygenation group. A definitive clinical trial of peripheral oxygen saturation targets is feasible in critically ill children.