BACKGROUND: Patients undergoing aortic arch replacement are at high risk for neurologic injury. This study compared two different established neuroprotective strategies in patients undergoing elective transverse hemiarch replacement. METHODS: Twenty patients undergoing hemiarch replacement were prospectively randomized to receive deep hypothermic circulatory arrest with retrograde cerebral perfusion (DHCA+RCP) or moderate hypothermic circulatory arrest with antegrade cerebral perfusion (MHCA+ACP). All patients received neurologist-adjudicated examinations and magnetic resonance imaging before discharge. The primary end point was a composite of stroke, transient ischemic attack, and magnetic resonance imaging-adjudicated injury. Secondary end points were transient neurologic dysfunction, and the National Institutes of Health Stroke Scale, and neurocognitive scores. RESULTS: Randomization resulted in 11 DHCA+RCP patients and 9 MHCA+ACP patients. There was no difference in cardiopulmonary bypass, cross-clamp, or circulatory arrest times. MHCA+ACP patients underwent circulatory arrest at significantly warmer temperatures (26.3° ± 1.8°C) than DHCA+RCP patients (19.9° ± 0.1°C, p < 0.0001). There were no deaths or renal failure in either group. There was 1 stroke in each group. National Institute of Health stroke scale scores and neurocognitive test results were equivalent. Diffusion-weighted magnetic resonance imaging demonstrated lesions in 100% (9 of 9) of MHCA+ACP patients compared with 45% (5 of 11) of DHCA+RCP patients (p < 0.01). MHCA+ACP patients had a significantly higher number of lesions than DHCA+RCP patients (p < 0.01). The primary end point was achieved in 100% of MHCA+ACP patients compared with 45% of DHCA+RCP patients (p < 0.01). CONCLUSIONS: Although there was no significant difference in clinically evident neurologic injury, this pilot study suggests that MHCA+ACP may be associated with a higher incidence of radiographic neurologic injury than DHCA+RCP in patients undergoing elective hemiarch replacement.