In bone marrow transplantation, high-dose busulfan is given p.o., usually every 6 h over 4 consecutive days. Since this repeated administration might alter busulfan disposition, fluctuations in busulfan plasma levels were studied over the 4-day treatment period in 21 children (median age, 5 years) with malignant solid tumors. In addition, urinary excretion of unchanged busulfan was measured every 6 h in 4 patients. Busulfan (37.5 mg/m2 for 16 doses) was given on an empty stomach at 12 p.m., 6 p.m., midnight, and 6 a.m. for 4 consecutive days, starting at 12 p.m. Trough plasma levels, i.e., concentration 6 h after each dose and just before the next one, and urinary excretion of busulfan were measured using a gas chromatography-mass spectrometry assay. Busulfan trough plasma levels exhibited a significant circadian rhythm with a higher mean level at 6 a.m. compared to that at 12 p.m., 6 p.m., and midnight. This rhythm was characterized by a double amplitude (mean +/- SD) of 42 +/- 14% and an acrophase (maximum) occurring at 5:48 a.m. +/- 115 min. In addition, once the steady state was reached, no decreasing trend was observed in any patient. Busulfan renal clearance proved to be low since only 5.4 +/- 1.2% of the given dose were excreted unchanged in urine. In the 4 patients studied, busulfan urinary excretion exhibited a significant circadian rhythm which was apparently linked to the physiological circadian rhythm in urinary output. Ten of 20 evaluable patients developed hepatic venoocclusive disease (HVOD). A significant circadian rhythm in the plasma level was found in both HVOD and non-HVOD patients with no difference between the two groups with regard to the 24-h mean, amplitude, or acrophase. Thus, the circadian changes in busulfan trough plasma levels observed at the steady state were not related to the occurrence of HVOD in these children with solid tumors. Moreover, since this rhythm was stable from day 2 to day 4, it should not compromise dose adjustment.