Oxygen Thresholds and Mortality During Extracorporeal Life Support in Adult Patients

Oxygen Thresholds and Mortality During Extracorporeal Life Support in Adult Patients

Munshi, L et al
Critical Care Medicine: December 2017 - Volume 45 - Issue 12 - p 1997–2005

Objectives: Extracorporeal life support can lead to rapid reversal of hypoxemia and shock; however, it can also result in varying degrees of hyperoxia. Recent data have suggested an association between hyperoxia and mortality; however, this conclusion has not been consistent across the literature. We evaluated the association between oxygenation thresholds and mortality in three cohorts of extracorporeal life support patients. 
Design: We performed a retrospective cohort study using the Extracorporeal Life Support Organization Registry. Setting: We evaluated the relationship between oxygenation measured 24 hours after extracorporeal membrane oxygenation onset and mortality (2010–2015). Patients: The extracorporeal life support cohorts were as follows: 1) veno-venous extracorporeal membrane oxygenation for respiratory failure, 2) veno-arterial extracorporeal membrane oxygenation for cardiogenic shock, and 3) extracorporeal cardiopulmonary resuscitation. Interventions: The relationships between hypoxemia (PaO2 < 60mm Hg), normoxia (PaO2 60–100mm Hg), moderate hyperoxia (PaO2 101–300mm Hg), extreme hyperoxia (PaO2 > 300 mm Hg), and mortality were evaluated across three extracorporeal life support cohorts. 
Measurements and Main Results: Seven hundred sixty-five patients underwent veno-venous extracorporeal membrane oxygenation, 775 patients underwent veno-arterial extracorporeal membrane oxygenation, and 412 underwent extracorporeal cardiopulmonary resuscitation. During veno-venous extracorporeal membrane oxygenation, hypoxemia (odds ratio, 1.68; 95% CI, 1.09–2.57) and moderate hyperoxia (odds ratio, 1.66; 95% CI, 1.11–2.50) were associated with increased mortality compared with normoxia. There was no association between oxygenation and mortality for veno-arterial extracorporeal membrane oxygenation. Moderate hyperoxia was associated with increased mortality during extracorporeal cardiopulmonary resuscitation compared with normoxia (odds ratio, 1.77; 95% CI, 1.03–3.30). An exploratory analysis did not find more specific PaO2 thresholds associated with mortality within moderate hyperoxia. 
Conclusions: Moderate hyperoxia was associated with increased mortality in patients undergoing veno-venous extracorporeal membrane oxygenation for respiratory failure and extracorporeal cardiopulmonary resuscitation. Hypoxemia was associated with an increased mortality in veno-venous extracorporeal membrane oxygenation. No association was seen between oxygenation and mortality in veno-arterial extracorporeal membrane oxygenation which may be due to early death driven by the underlying disease.