Abstract Body: Introduction: We present a novel demonstration of brain-directed management of venoarterial extracorporeal membrane oxygenation (ECMO) using real-time feedback from non-invasive neurometabolic optical monitoring (NOM) of the cerebral oxygen extraction fraction (OEF) in pediatric swine. In our prior work, OEF>0.82 was associated with hypoxic-ischemic brain injury during ECMO. Maintaining OEF below this threshold may prevent injury.

Hypothesis: Cerebral OEF levels of 0.8, 0.7, and 0.6 will be achieved during 7 h of brain-directed ECMO after prolonged cardiac arrest.

Methods: Continuous cerebral OEF was acquired in a pediatric swine model (sus scrofa, 8-12 kg) of ECMO-assisted cardiopulmonary resuscitation (CPR) using NOM on the left forehead. Animals had 8 min of ventricular fibrillation followed by 30 min of low-flow ECMO (20 ml/kg/min) with epinephrine (0.02 mg/kg) boluses every 4 min to simulate prolonged CPR. Then, ECMO flow rate was increased to 70-100 ml/kg/min to reflect clinical return of circulation. After 1 h, animals were randomized to a cerebral OEF target of 0.8 (n=11), 0.7 (n=12), or 0.6 (n=12) and brain-directed ECMO management was performed for 7 h; OEF was re-assessed every 5 min. If the OEF target was not achieved (±0.05 of target), first, ECMO flow rate was adjusted between 50-120 ml/kg/min, second, arterial blood pressure (ABP) was modulated using vasoactive infusions. Blood gas, hematocrit, anticoagulation, and anesthesia were otherwise standardized across all subjects. Mitochondrial respirometry assessed maximal oxidative phosphorylation of cortical tissue following ECMO as a metric of brain health which we have shown to correlate with neurologic outcomes across injury models. Associations with mean cerebral OEF, ECMO pump flow, and ABP were examined by linear regression.

Results: Brain-directed ECMO achieved and maintained respective OEF targets for 79% [57, 93] of the 7 h period (median [IQR] across animals). 71% (n=25/35) of animals achieved their target within 1 h of initiating brain-directed management. OEF, but not pump flow or ABP, was significantly correlated with maximal oxidative phosphorylation (p=0.005).

Conclusion: Our results demonstrate the efficacy and precision of real-time modulation of cerebral OEF during ECMO based on neurophysiologic feedback from the individual patient. The association of OEF with brain health suggests that brain-directed management may provide new opportunities to mitigate neurological injury.