A single masked trial of cyclosporin A 5 mg/kg/day versus monthly 1 g intravenous boluses of cyclophosphamide was conducted among 23 patients with Behçet's syndrome and active, potentially reversible uveitis. The trial was unmasked after a mean of 12 (SD 2) months for the cyclosporin A group (n = 12) and a mean of 10 (SD 3) months for the cyclophosphamide group (n = 11). During the initial 6 months the visual acuity significantly improved (p < 0.001) in the cyclosporin A group whereas this was not observed in the cyclophosphamide group. The subsequent follow-up of patients up to 24 months suggested that the initial improvement in visual acuity with cyclosporin A was not sustained. More extensive and especially long-term studies of cyclosporin A in the uveitis of Behçet's syndrome are warranted.