UK-ROX

Conservative Oxygen Therapy in Mechanically Ventilated Critically Ill Adult Patients: The UK-ROX Randomized Clinical Trial

What was it?

Multicenter, open-label, randomized clinical trial (UK-ROX) evaluating conservative oxygen therapy (target SpO2 88-92%) versus usual oxygen therapy (target SpO2 ≥96%) in mechanically ventilated adult patients in UK intensive care units (ICUs).

The Devil in the Details!

• 34 UK ICUs, 1,872 patients analyzed (934 conservative, 938 usual oxygen therapy).
• Adult patients (≥18 years) receiving invasive mechanical ventilation for ≤48 hours.
• Primary outcome: 90-day all-cause mortality.
• Secondary outcomes: ICU mortality, hospital mortality, duration of mechanical ventilation, organ support-free days, and adverse events (e.g., hypoxia, cardiovascular events).
• Conservative oxygen group aimed for SpO2 88-92% using lowest FiO2 possible; usual care targeted SpO2 ≥96%.
• Conducted from June 2018 to March 2022.

The Results!

• No significant difference in 90-day mortality: 38.0% (conservative) vs. 37.2% (usual) (adjusted hazard ratio 1.04 [95% CI 0.90-1.20], p=0.58).
• Conservative oxygen group had lower median FiO2 (0.30 vs. 0.40, p<0.001) and lower time-weighted average SpO2 (91.8% vs. 96.2%, p<0.001).
• No significant differences in secondary outcomes, including ICU mortality (31.1% vs. 30.5%, p=0.79), hospital mortality (40.2% vs. 39.8%, p=0.87), or organ support-free days (median 8 vs. 9, p=0.62).
• Adverse events: Conservative group had more hypoxic episodes (11.4% vs. 7.8%, p=0.01) but fewer hyperoxic episodes (2.1% vs. 15.3%, p<0.001).
• Subgroup analysis showed no consistent benefit or harm across patient types (e.g., sepsis, brain injury).

They concluded

The authors concluded that conservative oxygen therapy (SpO2 88-92%) did not significantly reduce 90-day mortality compared to usual oxygen therapy in mechanically ventilated ICU patients, with increased hypoxic episodes but reduced hyperoxia.

Gripe Point Summary!

Detailed gripes below, but key issues include:

• Open-label design risks bias in oxygen titration.
• Broad inclusion criteria may mask subgroup-specific effects.
• No long-term functional outcomes assessed.
• Higher hypoxic episodes in conservative group raise safety concerns.
• Limited generalizability outside UK ICU settings.
• Lack of blinding for clinical staff could influence co-interventions.
• Arbitrary SpO2 targets may not reflect optimal physiological range.
• Small effect size and wide confidence intervals limit definitive conclusions.

Our Summary

In mechanically ventilated ICU patients, conservative oxygen therapy (SpO2 88-92%) did not reduce 90-day mortality compared to usual care (SpO2 ≥96%), despite achieving lower oxygen exposure. Increased hypoxic episodes suggest caution, while reduced hyperoxia is a potential benefit. This large, well-conducted trial challenges the push for conservative oxygen strategies in all ICU patients, but further studies are needed to refine targets for specific populations.

Who’s worked on this before?

• ICU-ROX Investigators (2020): Conservative oxygen therapy in critically ill patients (New Zealand/Australia).
• Girardis et al. (2016): Effect of conservative vs. conventional oxygen therapy on mortality in ICU patients.
• Panwar et al. (2016): Conservative vs. liberal oxygenation targets in mechanically ventilated patients.
• Barrot et al. (LOCO2, 2020): Liberal vs. conservative oxygen in ARDS patients.
• Young et al. (HOT-ICU, 2021): Lower vs. higher oxygenation targets in acute hypoxemic respiratory failure.
• Schjørring et al. (2021): Conservative oxygen therapy in ICU patients with acute illness

Further gripes

• Open-label design: Lack of blinding for clinicians adjusting oxygen could introduce bias in FiO2 titration or co-interventions (e.g., ventilator settings, sedation).
• Broad inclusion criteria: Heterogeneous population (sepsis, trauma, neurological injury) may dilute effects in specific subgroups where conservative oxygen could be beneficial or harmful.
• Hypoxic episodes: 11.4% incidence in conservative group vs. 7.8% in usual care raises safety concerns, particularly for patients with brain injuries or cardiovascular disease.
• No functional outcomes: Focus on mortality and organ support misses long-term quality of life or neurological outcomes, critical in ICU survivors.
• SpO2 target rigidity: 88-92% vs. ≥96% may not reflect physiological optima; personalized targets based on patient condition (e.g., PaO2/FiO2 ratio) could be more relevant.
• Generalizability: UK ICU practices (e.g., staffing, protocols) may differ from other healthcare systems, limiting applicability.
• Statistical power: While large, the trial’s wide confidence intervals (hazard ratio 0.90-1.20) suggest it may be underpowered to detect small but clinically meaningful differences.
• Co-intervention variability: Uncontrolled differences in sedation, weaning protocols, or fluid management could confound outcomes.
• Short-term focus: No data on outcomes beyond 90 days, missing potential delayed effects of hyperoxia or hypoxia.
• Subgroup analysis limitations: Pre-specified subgroups (e.g., sepsis, brain injury) showed no consistent trends, but small sample sizes in some subgroups reduce confidence.

CCN’s Reflection

The UK-ROX trial is a robust, large-scale effort to settle the debate on oxygen targets in ICU patients, showing no clear mortality benefit for conservative oxygen therapy. Its multicenter design and rigorous methodology are strengths, but the open-label approach and increased hypoxic episodes in the conservative arm raise concerns. The lack of functional outcomes and broad patient inclusion leave questions about who might benefit from lower oxygen targets. This trial suggests a one-size-fits-all approach to oxygen therapy may not work—future research should focus on tailored strategies for specific ICU populations. For now, usual care remains a safe bet, but don’t hold your breath for a paradigm shift just yet!

Review by JW.