Study design: Randomized study. Objectives: To evaluate the effects of thalidomide on spinal cord ischemia/reperfusion injury via reduced TNF-a production. Setting: Animal experimental laboratory, Clinical Research Institute of Seoul National University Hospital, Seoul, Korea. Methods: Spinal cord ischemia was induced in rabbits by occluding the infrarenal aorta. Rabbits in group N did not undergo ischemic insult, but rabbits in groups C (the untreated group), THA, and THB underwent ischemic insult for 15 min. The THA and THB groups received thalidomide (20 mg/kg) intraperitoneally (i.p.) before ischemia, but only the THB group received thalidomide (i.p., 20 mg/kg) after 24 and 48 h of reperfusion. After evaluating neurologic functions at 1.5 h, 3, and 5 days of reperfusion, rabbits were killed for histopathologic examination and Western blot analysis of TNF-a. Results: The THA and THB groups showed significantly less neurologic dysfunction than the C group at 1.5 h, 3, and 5 days of reperfusion. The number of normal spinal motor neurons in ventral gray matter was higher in THA and THB than in C, but no difference was observed between THA and THB. Western blot analysis showed a significantly higher level of TNF-a in C than in THA and THB at 1.5 h of reperfusion, but no difference was observed between C, THA, or THB at 3 or 5 days of reperfusion. Conclusion: Thalidomide treatment before ischemic insult reduces early phase ischemia/ reperfusion injury of the spinal cord in rabbits.