Renal function was studied in 18 patients with Type 1 diabetes mellitus. All were participating in the Canadian-European randomized placebo-controlled cyclosporin trial in newly diagnosed Type 1 diabetic patients, nine being randomized to placebo, and nine to cyclosporin A. During treatment for 12 to 18 months, cyclosporin A caused significant reductions in the glomerular filtration rate (before drug withdrawal, cyclosporin 97 +/- 18 vs placebo 125 +/- 16 ml min-1 1.73-m-2, p less than 0.05), renal plasma flow (454 +/- 83 vs 536 +/- 70 ml min-1 1.73-m-2, p less than 0.05), and lithium clearance (17 +/- 3 vs 28 +/- 5 ml min-1 1.73-m-2, p less than 0.05). The fractional proximal reabsorption was increased (0.82 +/- 0.03 vs 0.78 +/- 0.03, p less than 0.05), and the fractional distal sodium reabsorption reduced (0.88 +/- 0.03 vs 0.94 +/- 0.02, p less than 0.05). These results are in accordance with the hypothesis that the nephrotoxic effect of cyclosporin A results from a preferential constriction of afferent glomerular vessels. One year after withdrawal of the drug, all variables were similar in the two groups, except for blood glucose control which was worse in the cyclosporin A treated group. When corrected for differences in blood glucose control it appeared that in three out of nine patients glomerular filtration rate had not completely returned to the reference range of the placebo group. We conclude that the nephrotoxic side-effects of cyclosporin A treatment for 1 year are reversible. There are, however, signs of minor and perhaps chronic renal injury.