Conservative versus liberal oxygenation targets for children admitted to PICU

In their pragmatic trial in 15 paediatric intensive care units (PICUs) in the UK, invasively ventilated children older than 38 weeks corrected gestational age and younger than 16 years without cyanotic heart disease, pulmonary hypertension, or acute brain injury who were admitted to PICUs as an emergency were randomly allocated to a conservative (peripheral oxygen saturation [SpO2] 88–92%) or liberal (SpO2 >94%) target for bedside clinician-driven ventilator titration. A meta-analysis of five neonatal trials, including nearly 5000 infants, found that aiming for SpO2 of 85–89% versus 91–95% increased mortality while reducing severe retinopathy, although only one individual trial showed increased mortality overall.3,4 Similarly, several trials of heterogeneous cohorts of critically ill adults have been done, and except for one single centre study comparing universal supplemental oxygen to drive SpO2 to 97% or greater versus as-needed oxygen targeting SpO2 94–98%, no large trial has shown improvement in mortality or other major outcomes with lower oxygen targets; one study of adults with acute respiratory distress syndrome even suggests harm.5–11 Overall, meta-analysis of 17 studies including more than 10 000 critically ill adults shows no effect of oxygenation targets on mortality, serious adverse events, or quality of life. 12 While critically ill children should not be considered either small adults or overgrown neonates, the lack of consistent benefit across trials limits enthusiasm for declaring SpO2 of 88–92% the new standard of PICU care based on one trial. [...]as in other trials, the conservative group was often above-range, with only a quarter of SpO2 values in goal range, and many children having deviations of 3 h or more out-of-range without ventilator adjustment. 6–9 Thus, it is impossible to conclude if achieving an SpO2 of 88–92% improves outcomes. [...]the risk that setting an aspirational goal of SpO2 88–92%—but really just wanting to avoid hyperoxia—could result in lax implementation of other patient goals (ie, targeted temperature management, chest compression rate, etc) must be considered.