To evaluate the efficacy and safety of leflunomide in the treatment of proliferative lupus nephritis, a prospective multi-centre observational study was conducted. Patients with biopsy proven proliferative lupus nephritis were assigned to receive either leflunomide or cyclophosphamide with concomitant prednisone. Leflunomide was given orally with a loading dose of 1 mg/kg/day for 3 days followed by 30 mg/day. Intravenous cyclophosphamide was administered monthly at a dose of 0.5 g/m2 of body-surface area. A total of 110 patients were enrolled, 70 in the leflunomide group and 40 in the cyclophosphamide group. The complete remission rate in the leflunomide group was 21% and partial remission rate 52%, as compared with 18% and 55%, respectively, in the cyclophosphamide group. Renal parameters and systemic lupus erythematosus disease activity index improved significantly and similarly in both groups. Serum creatinine decreased or stabilized in both treatment groups. No significant difference was noted with respect to clinical outcome between groups. Repeat biopsy also showed a significant reduction of active lesions in kidney pathology after 6 months of leflunomide treatment. Major adverse events, similar in both treatment groups, included infection, alopecia and hypertension. Leflunomide, compared with cyclophosphamide, in combination with prednisone was effective in the induction therapy of proliferative lupus nephritis and was generally well-tolerated.