Background: Targeted temperature management (TTM) at lower temperatures has shown variable neuroprotective benefits in clinical trials of cardiac arrest. We hypothesized that moderate hypothermia at the time of VA-ECMO reperfusion followed by early rewarming would reduce brain injury in a swine model of refractory VF, compared to mild hypothermia. Methods: Fourteen female pigs had electrically induced VF and conventional CPR for 30 minutes. Animals were cannulated to VA-ECMO and randomized to TTM to 34°C for 4 hours (mild hypothermia, n=7) or 24°C for 1 hour followed by early rewarming to 34°C over 3 hours (moderate hypothermia, n=7). Cooling was initiated upon ECMO reperfusion by circulating ice water through the oxygenator. Brain temperature and cerebral and systemic hemodynamics were continuously monitored. After four hours of post-arrest and ECMO care, brain tissue was obtained for examination. Groups were compared using paired t- and Wilcoxon rank sum tests. Results: The mean time to target brain temperature was 5 and 35 minutes in the mild and moderate groups, respectively (Panel, A). CPR hemodynamics were similar. Carotid blood flow (CBF) was higher in moderate hypothermia animals (79.0±1.56 vs 60.7±1.86% pre-arrest baseline, P <0.001) throughout the post-arrest period. There were no differences in ECMO flow, mean arterial pressure, vasopressor/fluid requirement, bleeding, or thromboelastography. No significant difference in histologic damage score was observed between groups (10.5 [7, 11.8] vs 13 [9, 20], with 24°C vs 34°C respectively P =0.25; Panel, B). Immunohistochemistry markers for white matter and neuronal injury also did not significantly diverge. Conclusion: In a swine model of refractory cardiac arrest treated with ECPR, moderate hypothermia initiated upon reperfusion followed by early rewarming did not appear to be neuroprotective. Histological brain injury at 4 hours post-resuscitation was not affected by the degree of hypothermia.