OBJECTIVE: To evaluate corticosteroid prescribing patterns in childhood-onset systemic lupus erythematosus (SLE), comparing four academic pediatric rheumatology practices. METHODS: Patients with childhood-onset SLE (n=72) treated at four large pediatric rheumatology centers were studied at 3-month intervals for 18 months. Information on medication use, disease activity as measured by the SLEDAI and the SLAM; and disease damage by the SLICC/ACR Damage Index was collected. RESULTS: At the time of enrollment, patients at each center were similar for disease duration, age, frequency of renal involvement and disease damage. Prednisone (mean 9 mg/day) was continued during 72% of periods of inactive disease for at least 3 months (SLEDAI=0). Centers differed in the use of intravenous pulse methylprednisolone (p<0.0001). Even when adjusted for between-center differences in patient weight, race and disease activity, centers also significantly differed in the dose of prednisone (p<0.05). The center with the largest between-patient variability in the dose of prednisone prescribed to its patients showed the smallest between-patient variance in patient disease activity. CONCLUSIONS: Corticosteroids are commonly used for the treatment of childhood-onset SLE, even when the disease is inactive. There appears to be important between-center differences in the use of intravenous and oral corticosteroids for childhood-onset SLE therapy that cannot be explained by patient disease activity corticosteroid prescribing patterns influence disease control. Further studies are needed to determine whether differences in practice patterns lead to significant differences in longer-term disease outcomes with childhood-onset SLE.