Thalidomide has been reported to be an effective agent for the treatment of chronic graft-vs.-host disease (GVHD). To determine its efficacy as a prophylactic agent for the prevention of chronic GVHD, a prospective randomized double-blind study was performed. A total of 59 patients were randomized to receive either placebo or thalidomide (200 mg orally twice a day) beginning 80 days after allogeneic bone marrow transplantation (BMT). Fifty-four evaluable patients were analyzed, 26 received placebo, and 28 received thalidomide. The characteristics of patients were well-balanced between the two groups. Following the first interim analysis conducted by the Data Safety Monitoring Board using an intent-to-treat approach, a statistically significant difference in the incidence of chronic GVHD was found. Patients receiving thalidomide developed chronic GVHD more often than patients receiving placebo (p = 0.06). Moreover, an apparent overall survival advantage was noted for patients receiving placebo compared to those receiving thalidomide (p = 0.006). Adjustment for possible confounding factors did not eliminate these negative effects of thalidomide. These results demonstrate that while thalidomide is an effective agent for the therapy of chronic GVHD, its use at the doses administered for the prophylaxis of chronic GVHD resulted in a paradoxical outcome with a higher incidence of chronic GVHD and a lower overall survival. We conclude that the early use of thalidomide results in a shift in the balance between GVHD and induction of tolerance. These data demonstrate again the importance of phase III double-blind controlled randomized studies.