On the basis that unbound concentration better correlates with response than total plasma or blood concentration, the inter- and intra-subject variability in the distribution of cyclosporin within blood and to plasma components was studied in renal transplant patients. Pharmacokinetic aspects were also studied. Blood samples were analysed from patients who received the drug both by a 72-h i.v. infusion and orally (7 mg.kg-1 twice daily). Steady-state was reached within 18 h of starting the i.v. infusion; the plasma data were best fitted by a biexponential equation with half-times of 0.13-1.02 h and 4.3-13.9 h, associated with the two phases. The mean plasma clearance was 700 ml/min. Concentrations during the infusions measured by RIA and HPLC were comparable. Oral profiles showed rapid and extensive absorption. The peak plasma concentrations were 1460-1880 micrograms.l-1 and occurred 2-4 h after dosing, with bioavailability estimates of 41-113%. Concentrations measured by RIA were higher than by HPLC. Blood-to-plasma concentration ratio measurements of cyclosporin at 37 degrees C decreased with increasing plasma concentration and increased with haematocrit. Fraction unbound, measured by ultra-centrifugation, was in the range 0.042-0.122 with an average of 0.068, and varied little in some patients but showed systematic changes with time in others. Cyclosporin binding was found to be related not only to the triglyceride but, more particularly, to the cholesterol-related lipoproteins in plasma. Monitoring cholesterol may be helpful in identifying patients with extremes in binding or with widely varying binding.