J Intensive Care Med. 2026 Sep 8:8850666261477389. doi: 10.1177/08850666261477389. Online ahead of print.
ABSTRACT
ObjectivesTo evaluate the association between early arterial hyperoxia and 28-day mortality in postcardiotomy VA-ECMO (PC-ECMO) adult patients.MethodsWe conducted a retrospective cohort study including 209 adults who received PC-ECMO across two tertiary centers between January 2019 and December 2024. The primary exposure was the 24-h mean arterial oxygen partial pressure (24h-mean PaO2) after ECMO support. The primary outcome was 28-day all-cause mortality. Cox proportional hazards models were used to assess associations between oxygen metrics and mortality, with adjustment for confounders.ResultsThe cohort had a median age of 59 years, 130 (62.2%) were male, and 80 (38.3%) received ECMO sopport due to failure to wean from cardiopulmonary bypass. The 28-day mortality was 54.1%. Across increasing 24h-mean PaO2 categories, 28-day survival declined from 64.1% in the <150 mm Hg group to 17.9% in the ≥300 mm Hg group (P < .001). Each 10 mm Hg increase in 24h-mean PaO2 was associated with a 2.8% increased mortality risk (adjusted HR 1.028, 95% CI: 1.011-1.046, P = .002). Severe hyperoxia (≥300 mm Hg) was independently associated with higher mortality (adjusted HR 2.392, 95% CI: 1.359-4.210). A nonlinear dose-response curve showed increased mortality beyond ∼200 mm Hg. Associations were more pronounced in elderly and male subgroups.ConclusionsEarly hyperoxia is independently associated with increased 28-day mortality in PC-ECMO patients in a dose-dependent manner, highlighting the need for conservative oxygen management.
PMID:42709762 | DOI:10.1177/08850666261477389