We have analyzed the plasma pharmacokinetics of busulfan in 272 patients receiving high-dose oral busulfan and intravenous cyclophosphamide in conjunction with allogeneic or autologous bone marrow transplantation. The patients ranged in age from 2 months to 59 yr (mean 10, median 12 yr) and had the following diagnoses: thalassemia or sickle cell anemia (n = 74); leukemia or myelodysplasia (n = 112); inborn errors of metabolism (n = 41) or immunodeficiency (n = 45). Plasma specimens were collected following the first dose for each patient which ranged from 1 to 4 mg/kg (mean +/- SD, 1.21 +/- 0.41, median 1.15). Busulfan was quantitated using ultraviolet absorbance detection after derivatization and HPLC separation. Pharmacokinetic parameters were derived by modeling the raw data to fit first-order single compartment kinetics. The kinetic parameters showed wide interpatient variability independent of age and diagnosis. There was a statistically significant correlation of age with the following parameters: area under the curve (AUC); maximal concentration; minimum concentration; clearance; volume of distribution and absorption half-time. The coefficients of determination (i.e. correlation coefficient squared) were low ranging from 0.04 to 0.12 implying only a small part (i.e. 4-12%) of the variance was explained by age. Although busulfan pharmacokinetics are age-related most of the variability is not explained by age or diagnosis.