Thalidomide shows antiangiogenic activity and it has been successfully employed in various tumors. Considering the poor therapeutic options for glioblastoma and the role of angiogenesis in malignant glioma cells growth, we investigated the therapeutic activity of thalidomide in patients affected by recurrent glioblastoma. Inclusion criteria were: recurrent glioblastoma pretreated with surgery and radiotherapy, age >/=18 years, adequate performance status, hematological, renal, and hepatic functions. Exclusion criteria included severe underlying diseases, neuropathy or concurrent radiotherapy. Eighteen patients entered the study, 17 of whom were assessable for toxicity and response. Most of patients were pretreated with chemotherapy (77.8%). Thalidomide was well tolerated: the most common side effects were constipation (76.5% of patients), somnolence (47%), and peripheral neuropathy (11.8%). One minimal response (MR) and 8 stable disease (SD) were observed, with an overall clinical benefit of 52.9%. Median time to progression and median overall survival (OS) for responders was 25 weeks (range 12-40) and 36 weeks (range 16-64), respectively. In conclusion, thalidomide induces modest side effects and it may be considered a valid therapeutic option for patients with recurrent glioblastoma.