Steroid-resistant (SR) forms of nephrotic syndrome (NS) have a poorer outcome in blacks compared to other racial groups. In this study, 223 children with SRNS, aged 1-16 years old, were analysed retrospectively for the period 1976-2004. Treatment schedules included oral cyclophosphamide (2-3 mg/kg) with prednisone 0.5-1 mg/kg (maximum 60 mg) only (n=90); prednisone on alternate days with methylprednisolone (30 mg/kg, maximum 1 g) and oral cyclophosphamide (n=117); oral prednisone on alternate days, three doses of intravenous methylprednisolone on alternate days and monthly doses of intravenous cyclophosphamide (500-750 mg m(-2) dose(-1)x7 doses monthly) (n=10); or cyclosporine 5 mg kg(-1) day(-1) adjusted to a trough level of 150-200 mg/ml (n=6). We compared the clinical and biochemical characteristics and outcome using different forms of therapies. A total of 183 (82.1%) underwent biopsy; 84 (45.9%) were Indian and 99 (54.1%) were black. Sixty-six (36.1%) had minimal change NS, 66 (36.1%) had focal segmental glomerulosclerosis (FSGS), 15 (8.2%) had a proliferative form of NS, and 36 (19.7%) had other forms of NS. Of the 84 Indian children biopsied, 58 (69.0%) were in complete remission, including 29 of 40 (72.5%) treated with oral cyclophosphamide and prednisone only. Of the 99 black children who were biopsied, 20 (20.2%) achieved complete remission; none of those treated with oral cyclophosphamide and prednisone only achieved complete remission. Of the 40 Indian children who were not biopsied who received only oral prednisone and cyclophosphamide, 32 (80%) achieved complete remission. This study shows Indian children with SRNS respond better to treatment than black children (69.0 vs. 20.2%). Since 80% of Indian children with SRNS responded to a trial of oral cyclophosphamide and prednisone, we propose the use of oral cyclophosphamide therapy in non-black children before embarking on renal biopsy.