BackgroundHyperoxia has been associated with adverse outcomes in post cardiac arrest (CA) patients. However, little data are available from mixed cohorts, where extracorporeal membrane oxygenation cardiopulmonary resuscitation (ECPR) and conventional CPR (CCPR) were utilised. The independence of effect of hyperoxia in this setting is not clear. Study-objective was to examine the association between hyperoxia and 30-day mortality in a mixed cohort of ECPR and CCPR patients.Methods and designThis was a retrospective cohort study of CA patients admitted to a tertiary level cardiac arrest centre in Australia from 1st January 2013 to 31st August 2018. Mean arterial oxygen levels (PaO2) and episodes of extreme hyperoxia (PaO2 ≥ 300mmHg) were analysed over the first 8 days. The primary outcome was 30-day mortality.ResultsA total of 169 post CA patients were assessed over a 6.5-year time period: 79 patients undergoing ECPR vs 90 patients undergoing CCPR. The mean age of the cohort was 54 (± 17) years; 126/169 (74%) were male and 119/169 (70%) were treated for out of hospital cardiac arrest (OHCA). Compared to CCPR, ECPR patients were younger, had a longer low flow time and higher illness severity scores on admission. Mean PaO2-levels were higher in patients in the ECPR vs CCPR group (211mmHg ± 58.4 vs 119mmHg ± 18.1; p < 0.0001) as was the proportion with at least one episode of extreme hyperoxia (58/79 (73%) vs 36/90 (40%), p < 0.01). ECPR patients presented with a higher mortality (54.4%) vs CCPR patients (34.4%). After adjusting for age, sex, BMI, highest lactate pre-treatment, use of ECMO, low flow time, pulse pressure on admission day, and severity of illness (APACHE III score), any episode of extreme hyperoxia was independently associated with a 2.57-fold increased risk of 30-day mortality (OR: 2.57, 95% CI: 1.09–6.06; p = 0.031) irrespective of the CPR-mode.ConclusionWe found extreme hyperoxia (PaO2-level ≥ 300mmHg) was more common in ECPR patients in the first 8 days post CA and was independently associated with higher 30-day mortality, irrespective of whether ECPR was employed. Prospective studies that compare different oxygen targets are needed to see if a strategy of lower oxygen exposure improves outcomes.