Systemic sclerosis (SSc) is a multisystem disease of presumed autoimmune pathogenesis for which no proven effective treatment exists. High-dose immunosuppressive therapy (HDIT) has been proposed as an investigational treatment for severe autoimmune diseases. Nineteen patients with poor-prognosis SSc underwent HDIT. The median age was 40 years (range, 23-61 years), the median modified Rodnan skin score (a measure of dermal sclerosis) was 31, and the median DLCO was 57%. Conditioning therapy involved 800 cGy total body irradiation (TBI) (+/- lung shielding to approximately 200 cGy), 120 mg/kg cyclophosphamide, and 90 mg/kg equine antithymocyte globulin. CD34-selected granulocyte-colony-stimulating factor-mobilized autologous blood stem cells provided hematopoietic rescue. With median follow-up at 14.7 months, the Kaplan-Meier estimated 2-year survival rate was 79%. Three patients died of treatment complications and one of disease progression. Two of the first 8 patients had fatal regimen-related pulmonary injury, a complication not found among 11 subsequent patients who received lung shielding for TBI. Overall, internal organ functions were stable to slightly worse after HDIT, and 4 patients had progressive or nonresponsive disease. As measured by modified Rodnan skin scores and modified health assessment questionnaire disability index (mHAQ-DI) scores, significant disease responses occurred in 12 of 12 patients evaluated at 1 year after HDIT. In conclusion, though important treatment-related toxicities occurred after HDIT for SSc, modifications of initial approaches appear to reduce treatment risks. Responses in skin and mHAQ-DI scores exceed those reported with other therapies, suggesting that HDIT is a promising new therapy for SSc that should be evaluated in prospective randomized studies.