Skip Ribbon Commands
Skip to main content
SharePoint

PublicComment

 Feedback

If you have any comments or questions on this page, please email us at:
 

Oxygen Dose After ROSC in Adults

Question Type:
Intervention
Full Question:
Among adults who have ROSC after cardiac arrest in any setting  (P), does  an inspired oxygen concentration titrated to oxygenation (normal oxygen saturation or partial pressure of oxygen)  (I), compared with  the use of 100% inspired oxygen concentration  (C), change  (O)?
Consensus on Science:
30% Versus 100% Inspired Oxygen for 60 Minutes After ROSC For the critical outcome of survival to hospital discharge with favorable neurologic outcome (CPC 1 or 2), 1 RCT enrolling 32 OHCA (of which 4 excluded) patients (very-low-quality evidence, downgraded for serious risk of bias, indirectness, and imprecision)(Kuisma 2006, 199) showed no difference between 30% inspired oxygen for 60 minutes after ROSC versus 100% inspired oxygen for 60 minutes after ROSC (8/14 versus 6/14; unadjusted RR for survival, 1.33; 95% CI, 0.63–2.84). For the critical outcome of survival to hospital discharge, 1 RCT (very-low-quality evidence, downgraded for small numbers, lack of blinding, indirectness, misallocation of patients)(Kuisma 2006, 199) showed no difference between 30% inspired oxygen for 60 minutes after ROSC and 100% inspired oxygen for 60 minutes after ROSC (10/14 versus 10/14; unadjusted RR for survival, 1.0; 95% CI, 0.63–1.60). Hyperoxia Versus Normoxia For the critical outcome of survival to 12 months with favorable neurologic outcome (CPC 1 or 2), 1 study(Vaahersalo 2014, 1463) of very-low-quality evidence (downgraded for very serious risk of bias and indirectness) showed no harmful effect associated with hyperoxia during the first 24 hours of ICU care. For the critical outcome of survival to hospital discharge with favorable neurologic outcome (CPC 1 or 2), 5 low-quality (downgraded as very serious bias and serious inconsistency, indirectness, confounding) observational studies showed conflicting results.(Kilgannon 2010, 2165; Bellomo 2011, R90; Janz 2012, 3135; Roberts 2013, 2107; Elmer 2015, 49) Two studies showed hyperoxia was worse than normoxia.(Kilgannon 2010, 2165; Janz 2012, 3135) Three studies reported favorable neurologic outcome as CPC 1 or 2. Very-low-quality evidence (downgraded because of very serious bias and serious inconsistency, indirectness, confounding) from a single-center study of 170 ICU patients treated with therapeutic hypothermia showed that the maximum PaO2 in the first 24 hours after arrest was associated with a worse outcome (poor neurologic status at hospital discharge; adjusted OR, 1.485; 95% CI, 1.032–2.136).(Janz 2012, 3135) Very-low-quality evidence (downgraded because of very serious bias and serious inconsistency, indirectness, confounding) from a single-center study of 193 ICU patients showed that the first PaO2 after ROSC was not associated with outcome (hyperoxia adjusted OR for poor neurologic outcome, 1.05; 95% CI, 0.45–2.42).(Roberts 2013, 2107) Very-low-quality evidence (downgraded because of very serious bias and serious inconsistency, indirectness, confounding) from a single-center study of 184 ICU patients showed that oxygen exposure over first 24 hours of ventilation was not associated with outcome with unadjusted and adjusted outcomes (effect size cannot be estimated from data).(Elmer 2015, 49) Two studies used surrogate measures of favorable neurologic outcome. Very-low-quality evidence (downgraded because of very serious bias and serious inconsistency, indirectness, confounding) from an observational study(Kilgannon 2010, 2165) showed worse independent functional survival at hospital discharge (hyperoxia versus normoxia, 124/1156 versus 245/1171 [29% versus 38%]; unadjusted OR, 0.45; 95% CI, 0.36–0.58). Very-low-quality evidence (downgraded because of very serious bias and serious inconsistency, indirectness, confounding) from an observational study(Bellomo 2011, R90) showed no difference in discharge to home (hyperoxia versus normoxia [27 versus 34%]; effect size cannot be estimated from data). For the critical outcome of survival to discharge (or survival to 30 days), very-low-quality evidence (downgraded because of very serious bias and serious inconsistency, indirectness, confounding) from 7 observational studies showed conflicting results.(Kilgannon 2010, 2165; Bellomo 2011, R90; Kilgannon 2011, 2717; Janz 2012, 3135; Ihle 2013, 186; Nelskyla 2013, 35; Elmer 2015, 49) Four studies showed hyperoxia worse than normoxia.(Kilgannon 2010, 2165; Kilgannon 2011, 2717; Janz 2012, 3135; Elmer 2015, 49) One study showed a worse outcome with hyperoxia versus normoxia based on the first ICU PaO2 (in-hospital mortality 63% versus 45%; adjusted OR hyperoxia exposure, 1.8; 95% CI, 1.5–2.2).(Kilgannon 2010, 2165) Another study showed a 100 mm Hg increase in PaO2 was associated with a 24% increase in mortality risk (OR, 1.24; 95% CI, 1.18–1.31).(Kilgannon 2011, 2717) One study showed no association between hyperoxia versus normoxia (based on the worse PaO2 in first 24 hours on ICU; adjusted OR for hospital mortality, 1.2; 95% CI, 1.0–1.5).(Bellomo 2011, R90) A single-center study of 170 ICU patients treated with therapeutic hypothermia documented that the maximum PaO2 in the first 24 hours after arrest was associated with a worse outcome.(Janz 2012, 3135) Survivors had lower maximum PaO2 (198 mm Hg; interquartile range, 152.5–282) versus nonsurvivors (254 mm Hg; interquartile range, 172–363); adjusted OR—higher PaO2 increased in-hospital mortality (OR, 1.439; 95% CI, 1.028–2.015). In a data linkage study of worse PaO2 (highest/lowest) in first 24 on ICU, hyperoxia was not associated with outcome (hospital mortality 47% versus 41%; adjusted OR hyperoxia versus normoxia, 1.2; 95% CI, 0.51–2.82).(Ihle 2013, 186) Another study of 122 ICU patients showed no difference between patients with hyperoxia (PaO2 greater than 300 mm Hg in first 24 hours after arrest) and normoxia (22/49 versus 25/70; unadjusted OR, 0.68; 95% CI, 0.32–1.44) for 30-day survival or survival to discharge (20/49 versus 24/70; unadjusted OR, 0.76; 95% CI, 0.36–1.61).(Nelskyla 2013, 35) In another study of 184 ICU patients, the 36% with severe hyperoxia had a mortality of 54%, and the presence of severe hyperoxia was associated with decreased survival in both unadjusted and adjusted analysis (adjusted OR for survival, 0.83 per hour exposure; 95% CI, 0.69–0.99).(Elmer 2015, 49) For the important outcome of survival to ICU discharge, very-low-quality evidence (downgraded because of very serious bias, serious indirectness, confounding) from 2 observational studies showed no harm from hyperoxia.(Ihle 2013, 186; Nelskyla 2013, 35) In a data linkage study of worse PaO2 (highest/lowest) in first 24 on ICU, hyperoxia was not associated with outcome (ICU mortality 35% versus 32% for hyperoxia versus normoxia; unadjusted OR, 1.16; 95% CI, 0.56–2.40).(Ihle 2013, 186-190) One observational study enrolling 122 ICU admissions patients showed no difference in survival to 30 days between patients with hyperoxia (PaO2 greater than 300 mm Hg in first 24 hours after arrest) and normoxia (ICU discharge 53% versus 46%; adjusted OR, 0.75; 95% CI, 0.36–1.55).(Nelskyla 2013, 35) Hypoxia Versus Normoxia For the critical outcome of survival to discharge (or survival to 30 days), very-low-quality evidence (downgraded because of very serious bias and serious indirectness, confounding) from 2 of 3 observational studies showed worse outcomes with hypoxia.(Kilgannon 2010, 2165; Bellomo 2011, R90; Ihle 2013, 186) One study showed a worse outcome with hypoxia versus normoxia based on the first ICU PaO2 (57% versus 45%; adjusted OR hypoxia exposure, 1.3; 95% CI, 1.1–1.5).(Kilgannon 2010, 2165) Another study documented that hypoxia versus normoxia (based on the worse PaO2 in first 24 hours on ICU) was associated with higher hospital mortality of 60% versus 47% (OR, 1.2; 95% CI, 1.1–1.4) but no difference in discharge to home (hypoxia/poor oxygen exchange versus normoxia 26% versus 24%).(Bellomo 2011, R90) In a data linkage study of worse PaO2 (highest/lowest) in first 24 hours on ICU, there was no difference in outcome between hypoxia and normoxia (for in-hospital mortality, 51% versus 41%; adjusted OR hypoxia versus normoxia, 0.93; 95% CI, 0.47–1.87).(Ihle 2013, 186) For the important outcome of survival to ICU discharge, very-low-quality evidence (downgraded because of very serious bias, serious indirectness, and confounding) from 1 observational study showed hypoxia was associated with a worse outcome.(Ihle 2013, 186) Worse PaO2 (highest/lowest) in first 24 hours in ICU was associated with a worse unadjusted outcome (ICU mortality 49% versus 32% for hypoxia versus normoxia; unadjusted OR, 2.15; 95% CI, 1.23–3.77; RR, 0.74; 95% CI, 0.56–0.96).
Treatment Recommendation:
We recommend avoiding hypoxia in adults with ROSC after cardiac arrest in any setting (strong recommendation, very-low-quality evidence). We suggest avoiding hyperoxia in adults with ROSC after cardiac arrest in any setting (weak recommendation, very-low-quality evidence). We suggest the use of 100% inspired oxygen until the arterial oxygen saturation or the partial pressure of arterial oxygen can be measured reliably in adults with ROSC after cardiac arrest in any setting (weak recommendation, very-low-quality evidence). Values, Preferences, and Task Force Insights In making these recommendations, we think, despite the very-low-quality evidence, there is likely to be far greater actual harm from hypoxia and, therefore, make a strong recommendation that hypoxia should be avoided. The evidence for harm associated with hyperoxia is of very low quality and inconsistent, hence the weak recommendation.
CoSTR Attachments:
CoS and TR for O2 after ROC_JS3Jan2015.docx    

 Contact Us

 
If you have any comments or questions on this page, please email us at: