Core Principle
- Both hypoxia and hyperoxia cause harm — avoid both extremes
- SpO2 is the preferred monitoring modality in routine critical care — treat as a vital sign
- Always specify upper and lower SpO2 limits (“swim between the flags”)
Default Targets
| Patient Group | SpO2 Target |
|---|---|
| Most ICU patients | 92–96% |
| COPD / chronic respiratory failure | 88–92% |
| Post-ROSC (cardiac arrest) | 94–98% |
| Bleomycin / paraquat toxicity | ~85% (O2 potentiates lung injury) |
| CO poisoning / decompression sickness | 100% high-flow (pending hyperbaric) |
SpO2 vs PaO2
Prefer SpO2 because:
- Non-invasive, continuous, immediate
- SaO2 is a better measure of systemic O2 delivery to tissues than PO2
- Reads within ±2% across the 70–100% range
- Chasing a “normal” PaO2 can lead to inappropriate over-oxygenation if SpO2 is already adequate
When SpO2 is unreliable → get ABG:
- Marked hypoxaemia
- Poor peripheral perfusion (unreliable trace)
- Dyshaemoglobinaemia (COHb, MetHb)
- Calculating P/F ratio (requires PaO2)
- Hyperoxia risk assessment (SpO2 stays 100% over a wide PaO2 range)
Hyperoxia
- Adverse effects on cardiovascular, pulmonary, CNS, and immune systems
- Mechanism: reactive oxygen species (ROS) + hyperoxia-induced vasoconstriction
- Historically liberal O2 has been ICU default — likely harmful
- Mixed evidence but signal of harm in: cardiac arrest, stroke, TBI, general ICU
Key trial: OXYGEN-ICU (Girardis 2016) — conservative O2 (SpO2 94–98%) vs liberal (≥97%) → ICU mortality 11.6% vs 20.2%. Significant methodological limitations — hypothesis-generating only.
Hypoxia
- Well-established harm across all critical illness groups
- Particularly dangerous in: TBI, post-cardiac arrest, stroke, ARDS survivors (long-term cognitive impairment)
- Risk of aiming too low: unplanned hypoxaemia if titration fails
Special Conditions
| Condition | Key Point |
|---|---|
| COPD | Target 88–92%; titrated O2 halved pre-hospital mortality vs high-flow in one RCT |
| ACS / Stroke | O2 only if hypoxaemic; supplemental O2 in non-hypoxaemic STEMI may increase infarct size |
| ARDS | No formal SpO2 guideline; subnormal oxygenation linked to long-term cognitive impairment |
| Post-cardiac arrest | 100% during CPR → titrate to 94–98% once ROSC established |
| Asthma / CAP | High-concentration O2 can cause CO2 retention — not unique to COPD |
Sources
Oxygen saturation targets in critical care • LITFL • CCC