Oxygen Targets in the ICU
Post-Cardiac Arrest
See ACLS - Post-Cardiac Arrest Care > Post-Arrest Oxygen Targets
Critically Ill Patients
| Study (Year) | Population | Intervention | Comparator | Outcome | Comment |
|---|---|---|---|---|---|
| AVOID (2015) | STEMI patients without hypoxia (SpO2 94%+) | Oxygen (8LPM facemask) | Oxygen only if SpO2 <94% | Trend towards decreased in-hospital mortality, but significant increase in recurrent MI (NNH=22) and arrhythmia (NNH=11) | Unblinded study, O2 not titrated in intervention. |
| OXYGEN-ICU (2016) | Mechanically ventilated adults, excluding AECOPD and severe ARDS | PaO2 70-100 or SpO2 94-96% | PaO2 < 150 or SpO2 97-100% | 11.6% vs 20.2% ICU mortality (p=0.01) | single centre study, unbalanced groups (favouring intervention arm) |
| ICU-ROX (2019) | Mechanically ventilated adults, excluding patients requiring avoidance/presence of hyperoxia, pregnancy | SpO2 91-96% | SpO2 > 90% and FiO2 > 0.30 | No difference in 28-day ventilator free days. No difference in 6-month mortality. | Largest such trial, better internal validity than OXYGEN-ICU |
| HOT-ICU (2021 NEJM) | Adults in the ICU on at least 10 LPM O2 or FiO2 > 0.50 (not necessarily intubated) | PaO2 target 60 mmHg | PaO2 target 90 mmHg | No differences in death at 90 days, life support dependence, shock, MI, stroke, intestinal ischemia. | |
| MEGA-ROX (PENDING) |
Takeaway
- Overall, the data suggest that for mechnically ventiled patients in the ICU, restrictive oxygenation targets may be beneficial (i.e. SpO2 91-96%) as compared to liberal targets (97%+), and we should use the minimal FiO2 required to achieve this. Hyperoxia may be harmful.
- As well, in patients with MI (not necessarily ventilated), avoid oxygen unless hypoxic.
ARDS
See Acute Respiratory Distress Syndrome > Oxygen Targets in ARDS