First line immunosuppressive treatment in steroid-resistant nephrotic syndrome in children is still open to discussion. We conducted a controlled multicentre randomized open label trial to test the efficacy and safety of cyclosporin A (CSA) versus cyclophosphamide pulses (CPH) in the initial therapy of children with newly diagnosed primary steroid-resistant nephrotic syndrome and histologically proven minimal change disease, focal segmental glomerulosclerosis or mesangial hypercellularity. Patients in the CSA group (n = 15) were initially treated with 150 mg/m(2) CSA orally to achieve trough levels of 120-180 ng/ml, while patients in the CPH group (n = 17) received CPH pulses (500 mg/m(2) per month intravenous). All patients were on alternate prednisone therapy. Patients with proteinuria >40 mg/m(2) per hour at 12 weeks of therapy were allocated to a non-responder protocol with high-dose CSA therapy or methylprednisolone pulses. At week 12, nine of the 15 (60%) CSA patients showed at least partial remission, evidences by a reduction of proteinuria <40 mg/h per m(2). In contrast, three of the 17 (17%) CPH patients responded (p < 0.05, intention-to-treat). Given these results, the study was stopped, in accordance with the protocol. After 24 weeks, complete remission was reached by two of the 15 (13%) CSA and one of the 17 (5%) CPH patients (p = n.s.). Partial remission was achieved by seven of the 15 (46%) CSA and two of the 15 (11%) CPH patients (p <0.05). Five patients in the CSA group and 14 patients in the CPH group were withdrawn from the study, most of them during the non-responder protocol. The number of adverse events was comparable between both groups. We conclude that CSA is more effective than CPH in inducing at least partial remission in steroid-resistant nephrotic syndrome in children.