Oxygen Targets in Comatose Survivors of Cardiac Arrest

Schmidt. NEJM 2022. DOI: 10.1056/NEJMoa2208686

Clinical Question

• In comatose patients following cardiac arrest does a restrictive compared with a liberal oxygenation strategy, reduce death or survival at discharge with a poor neurological outcome?

Background

• Hyperoxia can cause cellular damage via oxidative stress, absorption atelectasis, acute lung injury, central nervous system toxicity, reduced cardiac output and cerebral and coronary vasoconstriction, whereas hypoxia can cause ischaemic tissue injury
• Trials have compared restrictive vs liberal oxygen strategies in patients admitted to the ICU
• The OXYGEN-ICU study reported a significantly lower mortality in patients treated with a restrictive strategy
• Comparatively, in the larger ICU-ROX study there was no difference in mortality apart from in a post hoc analysis of a subgroup with ischaemic encephalopathy in whom a lower mortality was reported with conservative oxygen treatment
• Hypoxic brain injury is common following cardiac arrest. This trial therefore tried to determine if a restrictive oxygen strategy would benefit patients following cardiac arrest

Design

• Randomised controlled trial
• Randomisation using web-based system using variable block sizes that were stratified by site
• Non-blinded
• 2*2 factorial design
  • Patients also randomised to higher or lower MAP target, with results presented here
• Registered on clinicaltrials.gov
• Cognitive assessment tests completed via telephone interview
• 732 patients would provide 80% power to detect an absolute difference of 10% for the primary outcome assuming a mortality of 28% at a 2 sided alpha level of 0.05
• Intention to treat analysis
• PaO2 measured at pre-determined time points

Setting

• 2 tertiary cardiac arrest centres, Denmark
• Data collected March 2017 – December 2021

Population

• Inclusion:
  • Out-of-hospital cardiac arrest of presumed cardiac causes
  • Sustained ROSC >20 minutes
  • GCS <8
• Exclusion:
  • Presumed non-cardiac cause
  • >4 hours from ROSC to screening
  • SBP <80 despite medical optimisation
• 802 patients enrolled
  • 789 patients analysed in intention to treat analysis (consent declined in 12 and 1 patient randomised twice)
• Comparing baseline characteristics of restrictive vs. liberal group
  • Age: 62 vs 63
  • Male: 83% vs 79%
  • Medical history
    • Myocardial infarction: 23% vs 21%
    • Heart failure: 15% vs 20%
    • Stroke: 8% vs 7%
    • Chronic kidney disease: 5% vs 5%
  • Out-of-hospital arrest
    • Shockable rhythm: 85% vs 85%
    • Witnessed arrest: 85% vs 86%
    • Bystander CPR: 89% vs 86%
    • 1st defib by automated external defibrillator: 21% vs 26%
  • Finding/procedures at hospital arrival
    • pH: 7.21 vs 7.21
    • PaO2: 16.1 vs 17.1 kPa
    • Immediate coronary angiography: 92% vs 91%
    • Percutaneous coronary intervention: 40% vs 45%
  • Median time from cardiac arrest to randomisation 146 minutes (IQR 113-187)
  • Separation of oxygenation levels was seen within 2-4 hours and remained through the first 48 hours

Intervention

• Restrictive oxygenation
  • Target PaO2 of 9-10 kPa (68 -75 mmHg)
  • Initial FiO2 set at 0.3

Control

• Liberal oxygenation
  • Target PaO2 of 13-14 kPa (98 – 105 mmHg)
  • Initial FiO2 set at 0.6

Management common to both groups

• All patients underwent temperature control at 36C for at least 24 hours using surface cooling or intravenous cooling
• Target oxygenation interventions were initiated immediately after randomisation and maintained until extubation
• FiO2 adjusted to achieve target PaO2, but increased if SpO2 <93%
• Ventilator settings at discretion of treating physician

Outcome

• Primary outcome: Composite of death or discharge from hospital with severe disability or coma (Cerebral Performance Category of 3 or 4) within 90 days of randomisation – no significant difference
  • 32.0% in restrictive vs 33.9% in liberal
  • Hazard ratios 0.95 (95% CI 0.75-1.21), p=0.69
• Secondary outcomes:
• Comparing restrictive vs. liberal group
  • No significant difference in
    • Death within 90 days
      • 28.7% vs 31.1%
      • Hazard ratio 0.93 (95% CI 0.72-1.20)
    • Acute kidney injury with renal replacement therapy
      • 8.6% vs 11.9%
      • Hazard ratio 0.85 (95% CI 0.69-1.03)
    • Median CPC at 90 days: 1 vs 1
    • Median score on Montreal Cognitive Assessment at 90 days: 27 vs 27
    • Median neuron-specific enolase at 48 hours: 17 vs 18mcg/L
    • Adverse events
      • Infection: 26% vs 28%
      • Arrhythmia: 15% vs 13%
      • Uncontrolled bleeding: 4% vs 5%
• Sub-group analysis
  • Primary outcome consistent across pre-specified subgroups and no interaction with BP intervention

Authors’ Conclusions

• In comatose patients who had been resuscitated after out of hospital cardiac arrest, there was no difference between a restrictive and liberal oxygenation target with respect to in death or severe disability or coma at 90 days

Strengths

• Randomised controlled trial
• Intention to treat analysis
• Well balanced baseline characteristics
• PaO2 measured at pre-determined time points, and clear separation of PaO2 achieved

Weaknesses

• Open-label
• Study only performed in 2 centres in 1 country
• Patients in liberal oxygenation strategy treated with higher levels of PEEP
• Restrictive group had a mean PaO2 close to 10 kPa (Figure 1A), some even without additional oxygen supplementation
  • This was the upper limit of the restrictive group
• The PaO2 was ~16 and 17 kPa in both groups at randomisation
  • The authors comment that any suggestions that more aggressive or even pre-hospital interventions may have changed outcome are speculative
  • This would be supported by the recently published EXACT trial – this showed no difference in hospital survival when a SpO2 of 90 to 94% was compared to 98-100% in the pre-hospital setting
• Power calculation based on a reduction in the primary outcome from 28 to 18%

The Bottom Line

• In patients with out-of-hospital cardiac arrest targeting a PaO2 of 9-10 kPa compared with a PaO2 of 13-14 kPa did not affect mortality or likelihood of discharge from hospital with a poor neurological outcome

External Links

• article Oxygen Targets in Comatose Survivors of Cardiac Arrest

Metadata

Summary author: @davidslessor
Summary date: 21/10/2022
Peer-review editor: George Walker

Image by joakant from Pixabay