Background

Severe hyperoxia during venoarterial extracorporeal membrane oxygenation (VA-ECMO) has been associated with adverse clinical outcomes in observational studies. However, causal evidence and optimal oxygen targets remain uncertain. Hyperoxia may exacerbate oxidative stress and organ injury, particularly in the presence of systemic inflammation and ischemic shock. This study aimed to investigate the dose-dependent effects of membrane lung sweep oxygen fraction (FsO₂) during the early phase of VA-ECMO on systemic oxidative stress and early multi-organ injury markers, and to explore the mechanistic role of reactive oxygen species (ROS) using a rat VA-ECMO model.

Methods

In this randomized experimental study, Sprague–Dawley rats were divided into normal rats and rats with cardiogenic shock induced by septic cardiomyopathy using LPS. Graded membrane FsO₂ levels of 30%, 60%, and 90% were assigned, corresponding to normoxia, moderate hyperoxia, and severe hyperoxia, respectively. At least five animals were included in each endpoint-analysis subgroup. Specifically, nine animals were initially assigned to each septic cardiomyopathy subgroup to account for LPS-related mortality. A subset of animals received Tempol, a ROS scavenger. All animals underwent 3 h of VA-ECMO, followed by 3 h of post-weaning observation. Blood gas parameters and systemic oxidative stress markers were measured at multiple time points. At the experimental endpoint, blood and tissue samples were collected for analysis of oxidative stress markers, inflammatory cytokines, and histological injury in heart, lung, liver, and kidney tissues.

Results

Graded FsO₂ levels produced distinct and stable differences in arterial PaO₂ and SaO₂ during VA-ECMO support. In both normal and shock rats, severe hyperoxia (FsO₂ = 90%) caused greater systemic oxidative stress, metabolic acidosis and structural injury in the lung, liver, and kidney than normoxia (FsO₂ = 30%) or moderate hyperoxia (FsO₂ = 60%). In contrast, myocardial injury was not significantly modified by FsO₂ gradients. In animals without shock, these hyperoxia-associated changes were relatively mild and largely attenuated after VA-ECMO withdrawal, whereas shock markedly increased susceptibility to severe hyperoxia-induced early multi-organ injury. Notably, moderate hyperoxia (FsO₂ = 60%) was not associated with clearly worse oxidative stress or organ damage than normoxia (FsO₂ = 30%) in either normal or shock animals. Tempol administration attenuated oxidative stress and mitigated histological injury, and diminished differences among FsO₂ groups.

Conclusions

In this early-phase rat model of septic cardiomyopathy requiring VA-ECMO support, severe hyperoxia (FsO₂ = 90%) during VA-ECMO exacerbates metabolic disturbance and early multi-organ injury through ROS-mediated mechanisms. Moderate hyperoxia (FsO₂ = 60%) was not associated with clearly worse early injury markers than normoxia (FsO₂ = 30%) within the constraints of this short-duration preclinical model, suggesting that the upper boundary of oxygen exposure range during VA-ECMO in some contexts may be broader than previously assumed.