The delayed and incomplete sensory block of the sacral roots in epidural anesthesia often results from too huge the size of the first sacral nerve root. To overcome the problem, the enhancement of anesthetic effect on the sacral segment by injecting the local anesthetic toward the caudal direction has been reported. Thus, in this study, we compared the anesthetic effects of epidural anesthesia by catheterization toward caudal direction with that toward cephalic direction on the areas innervated by the sacral nerve roots.The study enrolled 45 ASA physical status I or II patients scheduled for ankle surgery or hemorrhoidectomy who were randomly and evenly divided into two groups. The epidural catheter was placed either cephalad (EU group) or caudad (ED group) and confirmed by C-arm fluoroscopy after catheter insertion. Each group finally consisted of 10 patients for ankle surgery and 10 for hemorrhoidectomy after exclusion of the ineligibles. The onset time, duration of, and recovery time from epidural anesthesia, as well as demand of intra-operative analgesic in the two groups were recorded and compared.No significant differences were noted in the patients' characteristics between the two groups. There were also no statistically significant differences in onset, duration, recovery time, and anesthetic level attained between the two groups. More patients in the EU group demanded analgesics during surgery than did patients in the ED group. However, there were no statistically significant differences in the rate of analgesic demand between the two groups.Epidural injection of local anesthetic via catheter oriented caudad or cephalad did not change the characteristics of epidural anesthesia including onset time, duration, anesthetic level, and analgesic effect on the surgical areas innervated by the sacral roots.