FK506 trough levels were measured by ELISA in paired whole-blood and plasma samples in 59 liver transplant recipients. Patients with nephrotoxicity had higher FK506 whole-blood and plasma levels (27.5 +/- 3.2 ng/ml and 1.44 +/- 0.14 ng/ml) than patients with stable liver function (15.2 +/- 2.1 ng/ml and 0.98 +/- 0.15 ng/ml, P < 0.05 and P < 0.01, respectively). Patients with acute rejection had FK506 whole-blood and plasma levels within the same range as patients with stable liver function. Patients with severe neurotoxicity had significantly higher FK506 whole-blood and plasma levels (31.3 +/- 6.8 ng/ml and 3.9 +/- 1.4 ng/ml) in comparison with patients with mild-to-moderate neurotoxicity (18.1 +/- 2.4 ng/ml and 1.1 +/- 0.13 ng/ml) (P = 0.048 and P < 0.001, respectively). Long-term use of FK506 was associated with a significant reduction in glomerular filtration rate at 1-year posttransplant in patients on primary FK506 treatment (33%, P < 0.001). The reduction in glomerular filtration rate correlated with the yearly mean FK506 plasma but not with whole-blood levels or FK506 dose. There was a correlation between FK506 whole-blood and plasma levels (r = 0.713, P < 0.001) but not between the levels (whole blood or plasma) and FK506 dose (mg/day or mg/kg/day). The mean FK506 whole-blood and plasma levels were 14.1 +/- 0.26 ng/ml and 0.96 +/- 0.75 ng/ml, respectively. There was a large intra- and interpatient variability in the ratio between whole-blood and plasma levels (range 1.0-73.5), with a mean ratio of 18.0 +/- 0.28 (+/- SEM). In conclusion, monitoring of FK506 trough levels is of importance to avoid nephro- and neurotoxicity, but monitoring is only of limited help to avoid acute rejection. Monitoring of FK506 levels in plasma seems to be superior to that in whole blood.