The influence of assay method on steady-state cyclosporine (CsA) pharmacokinetics was studied in 18 patients with leukemia or aplastic anemia undergoing allogeneic marrow transplantation who received i.v. CsA for prophylaxis or treatment of graft-versus-host disease. Since CsA is extensively metabolized and subject to biliary elimination, CsA courses were divided according to the presence (serum bilirubin greater than 2.0 mg/dl) or absence (serum bilirubin less than or equal to 1.2 mg/dl) of hepatic dysfunction. All patients had normal renal function. Serum CsA concentrations were measured concurrently by radioimmunoassay (RIA) and high-performance liquid chromatography (HPLC). Serum concentration-time data were analyzed by standard nonlinear regression methods. Systemic clearance (Cls) calculated from HPLC results was higher than that derived from RIA results, regardless of hepatic function (P less than 0.05). These data indicate that results obtained by RIA overestimate CSP concentrations, presumably because of the presence of crossreactive CSP metabolites. These differences can significantly affect derived pharmacokinetic parameters. Therefore, clinicians who make dosage recommendations based on pharmacokinetic parameters should consider the effect of assay method.