Background Retrograde cerebral perfusion (RCP) is used in some centers during aortic arch surgery for brain protection during hypothermic circulatory arrest. It is still unclear however whether RCP provides adequate microcirculatory blood flow at a capillary level. We used intravital microscopy to directly visualize the cerebral capillary blood flow in a piglet model of RCP. Methods Twelve pigs (weight 9.7 ± 0.9 kg) were divided into two groups (n = 6 each): deep hypothermic circulatory arrest (DHCA) and RCP. After the creation of a window over the parietal cerebral cortex, pigs underwent 10 minutes of normothermic bypass and 40 minutes of cooling to 15°C on cardiopulmonary bypass ([CPB] pH-stat, hemocrit 30%, pump flow 100 mL · kg −1 · min −1). This was followed by 45 minutes of DHCA and rewarming on CPB to 37°C. In the RCP group the brain was retrogradely perfused (pump flow 30 mL · kg −1 · min −1) during DHCA through the superior vena cava after inferior vena cava occlusion. Plasma was labeled with fluorescein-isothiocyanate-dextran for assessing microvascular diameter and functional capillary density (FCD), defined as total length of erythrocyte-perfused capillaries per observation area. Cerebral tissue oxygenation was determined by nicotinamide adenine dinucleotide hydrogen (NADH) autofluorescence, which increases during tissue ischemia. Results During normothermic and hypothermic antegrade cerebral perfusion the FCD did not significantly change from base line (97% ± 14% and 96% ± 12%, respectively). During retrograde cerebral perfusion the FCD decreased highly significantly to 2% ± 2% of base line values ( p < 0.001). Thus there was no evidence of significant capillary blood flow during retrograde cerebral perfusion. The microvascular diameter of cerebral arterioles that were slowly perfused significantly decreased to 27% ± 6% of base line levels during RCP. NADH fluorescence progressively and significantly increased during RCP, indicating poorer tissue oxygenation. At the end of retrograde cerebral perfusion there was macroscopic evidence of significant brain edema. Conclusions RCP does not provide adequate cerebral capillary blood flow and does not prevent cerebral ischemia. Prolonged RCP induces brain edema. However, there might be a role for a short period of RCP to remove air and debris from the cerebral circulation after DHCA because retrograde flow could be detected in cerebral arterioles.