Sixty children, with biopsy diagnosed focal segmental glomerulosclerosis (FSGS) and with unremitting nephrotic syndrome despite intensive therapy with adrenocortical steroids, were randomly allocated into a clinical trial comparing prednisone, 40 mg/m2 on alternate days for a period of 12 months (control group), with the same prednisone regimen plus a 90-day course of daily cyclophosphamide, 2.5 mg/kg in a single morning dose (experimental group). One-quarter of the children in each group had complete resolution of proteinuria. The proportions of children with increased, unchanged, and decreased proteinuria by the end of the study were the same in the two groups. Treatment failure was defined as an increase in serum creatinine of 30% or more or greater than 0.4 mg/dl, or onset of renal failure. Treatment failure occurred in 36% of the control group and 57% of the experimental group (P > 0.1). Five patients died during the trial, 3 in the experimental group and 2 in the control group. A Kaplan-Meier survival analysis revealed no significant differences between the two groups. Cyclophosphamide therapy for children with steroid-resistant FSGS is not recommended.