In juvenile IDDM patients, immunosuppression with cyclosporin A allows partial beta-cell function recovery and transient remissions of insulin dependency. The effects of this therapeutic approach, however, have not been evaluated in the long-term, since no reported trial exceeded 1 year. Here we analyze 130 diabetic children followed at our institution during the first years of their disease. Cyclosporin was given to 83 of them at an initial dose of 7.2 +/- 0.1 mg.kg-1.day-1, which was decreased stepwise then interrupted after 6-62 months, depending on the response to therapy. A total of 47 diabetic children, who served as control subjects in two trials, were pooled for comparison. Over 4 years, the cyclosporin-treated group kept plasma C-peptide approximately twice as high as the control group (P < 0.02). It took 5.8 +/- 0.6 years for C-peptide secretion stimulated by glucagon to become undetectable in the cyclosporin group versus 3.2 +/- 0.6 years in the control group (P < 0.02). Average insulin dose remained lower by 0.2-0.4 U.kg-1.day-1 and glycated hemoglobin by approximately 1% in cyclosporin-treated patients (P < 0.02), who also had less hypoglycemia than the diabetic control subjects (P < 0.05). After 4 years, differences between the groups became nonsignificant. We observed no significant secondary effects of cyclosporin. In conclusion, positive effects of low-dose cyclosporin in recently diagnosed clinical IDDM patients are prolonged beyond interruption of the drug. The magnitude and duration of the benefit, however, do not appear sufficient to justify this immunosuppressive treatment in clinical practice.