Epidural opioids must redistribute to the spinal cord to produce analgesia. It has been suggested that opioids may reach the spinal cord via the radicular arterial blood supply. This study was designed to test this hypothesis. Microdialysis probes were placed in the lumbar and thoracic spinal cord of six anesthetized pigs and a ligature placed around the thoracic aorta. At t = 0, fentanyl (300 micrograms) was injected into the epidural space and the spinal cord and plasma were sampled over 90 min to determine resultant spinal cord dialysate fentanyl concentrations. At t = 90, the thoracic aorta was occluded to eliminate distal radicular artery blood flow and fentanyl was again injected into the epidural space. The spinal cord was sampled for an additional 90 min to determine spinal cord dialysate fentanyl concentrations. At t = 90 min the aortic cross-clamp was released and spinal cord samples were collected for another 30 min. There was no difference in lumbar spinal cord dialysate concentrations of fentanyl between the control and cross-clamp conditions over the first 20 min after epidural administration. Between 20 and 90 min, lumbar spinal cord fentanyl concentrations were significantly greater during aortic cross-clamp. Fentanyl was not detected in the thoracic cord. Radicular artery blood supply does not redistribute fentanyl from the epidural space to the spinal cord in this model. However, radicular artery blood flow does play a role in clearing fentanyl from the spinal cord.