ECMO-facilitated rapid and deep hypothermia reduces brain injury on MRI following prolonged cardiac arrest in a translational swine model.

Introduction: Neuroprotective interventions after cardiac arrest are essential but largely lack evidence of efficacy. Early therapeutic hypothermia (TH) is the only intervention that has shown promise in humans. However, despite a consistent signal for efficacy in animal models, conflicting clinical data hamper clinical acceptance. Two potential causes for the lack of translation from animal studies to humans are the time to achieve target temperature in humans and the inability to cool to deep hypothermic states due to the inherent detrimental cardiac effects accompanying deep hypothermia. Given the observed inconsistent impact of TH on human patients with cardiac arrest despite animal data, we developed a perfusion-controlled, translational swine model to quantify the effects of rapid deep TH on HIBI, quantifying severity using magnetic resonance imaging (MRI) with diffusion-weighted imaging (DWI) at a controlled time threshold. Methods: Ten swine underwent cardiac arrest with 20 min of "no-flow" state, followed by resuscitation and controlled reperfusion using extracorporeal membrane oxygenation (ECMO). Animals were randomized to either control (normal temperature reperfusion) or rapid hypothermic reperfusion (RHR) (29 °C through ECMO-facilitated cooling). All swine underwent brain MRI with Diffusion Weighted Imaging (DWI) before cardiac arrest and then 2 h after ECMO reperfusion. Whole-brain gray and white matter apparent diffusion coefficient (ADC) values were compared pre- and post-ECMO cannulation and arrest in all animals. Results: At 45 min post-reperfusion, the mean temperature for RHR animals was 30.4 °C (95 % CI 29.6-31.1 °C), while for control animals it was 35.7 °C (95 % CI 34.9-36.5 °C, p < 0.0001). Whole brain ADC in RHR swine increased by a mean of 1.36 ± 4.09 %, while in control swine it decreased by a mean of 4.36 ± 4.50 % (Median difference of -5.91, 95 %CI -12.13 to -0.15; P value = 0.047). Discussion: Swine with induced cardiac arrest who underwent rapid ECMO-mediated cooling post-arrest had less cerebral hypoxic cellular injury, as quantified by changes on MRI DWI, than controls. These findings support the protective effect on neurologic injury of a rapid and brief period of induced deep hypothermia after cardiac arrest. Compared to prior translational models, our use of ECMO has the advantage of an ability to control important factors such as no-flow ischemic time and variability in post-arrest cardiac output as well as to mitigate complications of cardiac dysrhythmias that tend to arise from deep hypothermia. This portends a greater promise for translational success of ECMO-facilitated rapid cooling and potentially other ECMO-mediated models of cardiac arrest than experienced by previous attempts.