Autologous transplants for severe and refractory multiple sclerosis (MS) were proposed in 1997 and have been performed on about 200 selected patients worldwide. Phase I/II clinical studies have shown that high-dose immunosuppressive therapy suppresses inflammation in the CNS and may delay the progression of clinical disease. The procedure is associated with toxicity from the high-dose cytotoxic therapy and a risk of serious infections. There is a transplant-related mortality risk of 1-5%, requiring careful patient selection before transplantation. Treatment should be reserved for patients who have a significant chance of response, i.e. young patients with low disability scores but rapidly progressing disease who have inflammatory rather than neurodegenerative changes in the CNS. The long term effect of high-dose immunosuppression after transplantation on the frequency of relapse or progression of MS is unclear, but the initial concept of immune ablation by high-dose therapy and the reconstitution of normal immunity and tolerance from transplant-derived lymphocyte progenitors has given way to the concept of 'resetting' the immune system. The clinical effect of transplantation remains to be demonstrated in comparative studies.