Improved care and drug treatment have considerably decreased morbidity and mortality in childhood systemic lupus erythematosus (SLE), although renal failure, infections, and drug-related toxicity remain major concerns. In recent years, potent immunosuppressive drug regimens have been introduced to control the disease (1). Nevertheless, a small proportion of children have not responded to these drugs or have developed severe side effects. Autologous stem cell transplantation (ASCT) has recently been described as a possible treatment for severe chronic autoimmune diseases that have been refractory to conventional therapy (2, 3). This was based on several case reports that described prolonged remission of autoimmune disease in patients who were treated with bone marrow transplantation for a malignancy (4). Regression of adjuvant-induced arthritis was observed in rats after total-body irradiation (TBI) and autologous bone marrow transplantation (5, 6). The rationale of this therapy is to suppress the immune system maximally and then “rescue” the patient from prolonged or persistent aplasia by infusing autologous stem cells. Between 1994 and 1999, more than 20 cases of adult SLE patients treated with ASCT have been reported (7, 8), and participants in several workshops have proposed inclusion criteria for this experimental treatment. Herein we present the first report of ASCT in children with SLE. ASCT was administered to 2 children, both of whom were steroid dependent but experienced severe side effects from the steroids. They were treated with high-dose cyclophosphamide (CYC), which induced strong immunosuppression but little immunoablation. We added low-dose TBI to their regimens based on the finding, in an animal arthritis model, that this induces a very profound depletion of T memory cells (6). T cell depletion of the graft was performed to avoid reinfusion of putative autoreactive T cells. We hypothesize that immune reconstitution in the absence of mature T memory cells in a patient with aplasia may induce tolerance to unknown autoantigens responsible for the disease. The protocol was approved by our institutional research board. Written informed consent was obtained. Clinical details on the patients are shown in Table 1. Patient 1, a 14-year-old girl, presented at the age of 10 with fever, Raynaud’s phenomenon, malar rash, sunlight hypersensitivity, vasculitic skin lesions on her fingers and toes, and polyarthritis. A few months later she developed hemorrhagic pneumonitis, glomerulonephritis (World Health Organization [WHO] class IV), and pancytopenia. She was treated for 6 months with pulse CYC, which was stopped because of severe hemorrhagic cystitis. In the following years, azathioprine (AZA) and hydroxychloroquine (HCQ) were added to her treatment regimen. Tapering of the prednisone to 1 mg/kg/day resulted repeatedly in exacerbations of her SLE, with vasculitic lesions in the fingers, malar rash, fever, malaise, and a concomitant decrease in complement levels and increase in the anti–double-stranded DNA (anti-dsDNA) titer.

Patient 2, a 14-year-old boy, was diagnosed with SLE at the age of 9, presenting with Coombs-positive hemolytic anemia, polyarthritis, lymphadenopathy, and malar rash. A few months later, he lost vision in his right eye due to vasculitis. In the following years, he developed glomerulonephritis (WHO class IV) that remained active despite treatment with steroids (both pulse methylprednisolone and oral maintenance therapy), pulse CYC, AZA, and HCQ. He developed severe pericarditis that responded well to steroids. Tapering of steroids to 0.75 mg/kg/day resulted in flares of the pericarditis …