In 69 renal allograft recipients the highest-tolerated dose was given with respect to clinical events but without respect to the CsA plasma level (CsA-PL). The CsA dose was gradually decreased during the first 6-12 months after transplantation, and in some patients even later. The CsA dose after 12 months was 5-8 mg/kg/day and after 18 months 4-6 mg/kg/day, resulting in CsA-PL of 85-140 ng/ml and less than 50-110 ng/ml, respectively. CsA side-effects were usually seen in patients with high CsA-PL, but they were also encountered at levels normally seen in patients without toxicity. In the individual patient, acute CsA nephrotoxicity was associated with a significant rise in CsA-PL. In patients with acute nephrotoxicity a reduction of the CsA dose (mean 24%) was necessary to regain satisfactory renal function. All patients with several consecutive CsA-PL above 1000 ng/ml had hepatotoxicity or nephrotoxicity, or both, associated with severe morbidity and mortality. No difference was found between CsA-PL during acute rejection and during good renal function. The percentage of CsA determinations resulting in plasma levels below 50 ng/ml (the limit of detection) increased with the time of therapy and constituted 22% of all CsA-PL after 6 months of therapy. No rejections were seen later than 5 months after transplantation despite the low CsA-PL in many long-term treated patients. Treatment with high doses of methylprednisolone increased CsA-PL by 223%. Trimethoprimsulphamethoxazole in CsA-treated patients caused increases in serum creatinine levels. Monitoring of trough CsA plasma levels is recommended as a complement to clinical judgment. To avoid most nephrotoxicity and hepatotoxicity we have decided to keep the CsA-PL below 500 ng/ml during the first month, below 250 ng/ml the second month after transplantation, and below 200 ng/ml the third month after transplantation--and in long-term treated patients we now keep the CsA-PL between less than 50 and 150 ng/ml.