Conservative versus liberal oxygenation targets in critically ill children: the Oxy-PICU RCT

The optimal target for systemic oxygenation in critically ill children is unknown. Liberal oxygenation is widely practised but is associated with harm in observational studies.

To evaluate the clinical and cost-effectiveness of a conservative oxygenation target of peripheral oxygen saturation 88-92% compared with peripheral oxygen saturation > 94% in critically ill children admitted to paediatric intensive care unit as an emergency.

A pragmatic, open, multicentre, parallel-group, randomised clinical trial conducted in 15 National Health Service paediatric intensive care units and associated emergency transport services across England and Scotland.

Children aged > 38 weeks corrected gestational age and < 16 years, enrolled within 6 hours of being accepted for admission to paediatric intensive care unit as an emergency; receiving invasive mechanical ventilation with supplemental oxygen; and in face-to-face contact with paediatric intensive care unit or emergency transport services staff.

Adjustment of ventilator and inspired oxygen settings aiming to achieve peripheral oxygen saturation 88-92% (conservative oxygenation) or peripheral oxygen saturation > 94% (liberal oxygenation) during invasive mechanical ventilation.

Primary outcomes: duration of organ support at 30 days, with death by day 30 ranked as the worst outcome (clinical effectiveness) and incremental costs, quality-adjusted life-years and net monetary benefit at 12 months (cost-effectiveness). Secondary outcomes: incremental costs at 30 days; mortality at paediatric intensive care unit discharge, 30 days, 90 days and 12 months; time to liberation from ventilation; duration of organ support; length of paediatric intensive care unit and hospital stay; functional status at paediatric intensive care unit discharge; and health-related quality of life at 12 months.

Two thousand and forty children were randomised between 1 September 2020 and 15 May 2022. Consent was obtained for 1872 (94%) - 939 to the conservative and 933 to the liberal oxygenation group - who were included in the primary analysis. Duration of organ support or death in the first 30 days was lower in the conservative oxygenation group [probabilistic index 0.53, 95% confidence interval 0.50 to 0.55; p = 0.04 Wilcoxon rank-sum test, adjusted odds ratio 0.84 (95% confidence interval 0.72 to 0.99)]. Both components of the composite primary outcome and secondary outcomes favoured conservative oxygenation. Average costs at 30 days strongly indicated lower costs with conservative oxygenation. Longer-term estimated incremental costs and quality-adjusted life-years were lower and net monetary benefit marginally favoured conservative oxygenation but with wide uncertainty [incremental costs -£879 (95% confidence interval -9036 to 7278); quality-adjusted life-years 0.001 (-0.010 to 0.011); net monetary benefit £894 (95% confidence interval -7290 to 9078)].

Exclusion of two large paediatric intensive care unit populations, due to a lack of equipoise and the number of participants excluded because of not being able to obtain deferred consent.

Future work should focus on identification of the mechanisms underlying the observed benefit; trials of intermediate or lower peripheral oxygen saturation values in individuals at higher risk; and identification of individualised treatment effects in relation to oxygen therapy.

A conservative oxygenation target resulted in a greater probability of a better outcome in terms of duration of organ support at 30 days or death. Longer-term survival and health-related quality of life were consistent with the primary outcome. While conservative oxygenation is likely to reduce costs in the short term, longer-term cost-effectiveness was surrounded with wide uncertainty.

This synopsis presents independent research funded by the National Institute for Health and Care Research (NIHR) Health Technology Assessment programme as award number NIHR127547.