A 1 year, prospective multicentre study was performed to investigate the efficacy and safety of intermittent treatment with cyclosporin in psoriasis vulgaris. Subjects received cyclosporin (Neoral) 5 mg/kg per day until achieving 90% reduction in area affected, or for a maximum of 12 weeks. Those failing to demonstrate a satisfactory response were withdrawn. When further treatment was required, cyclosporin was recommenced. This cycle was repeated up to three times. Psoriasis activity was recorded using the area affected and sign scores for erythema, scaling and infiltration. Overall assessments of response and tolerability were recorded. Forty-one subjects, mean age 36, mean PASI 12.8, entered the first treatment period. Thirty-three received a second period of treatment and 16 a third. Eighteen failed to complete the study as planned: five were withdrawn due to adverse events, four due to treatment failure and nine due to protocol violations. At the end of each treatment period, significant improvements were seen in all efficacy parameters. Overall response was graded as 'considerable improvement' or 'minimal or no symptoms', by over 80% of subjects and investigators. Median intervals to relapse for subjects remaining in the study were 72 days (range 28-329) and 53 days (range 14-141) after periods 1 and 2, respectively. There were significant increases in mean serum creatinine and blood pressure during each treatment period. However, there were no significant differences in either parameter between baseline and the final follow-up visit. At the end of each treatment period, overall tolerability of the treatment was considered 'good' or 'very good' by over 80% of subjects and investigators.