The pharmacokinetics and metabolism of doxorubicin (DX) and epirubicin (epiDX) were investigated in eight cancer patients who received 60 mg/m2 of both drugs independently by intravenous (i.v.) bolus at 3-week intervals according to a balanced cross-over design. Unchanged DX and epiDX plasma levels followed a triexponential decay. Half-lives (t/2) of the three decay phases were longer for DX (t/2 alpha: 4.8 vs. 3 min; t/2 beta 2.57 h vs. 1.09 h; t/2 gamma 48.4 vs. 31.2 h). According to a model-independent analysis, the different plasma disposition kinetics of the two compounds appears to be related to a higher plasma clearance (PlCl) and to a lower mean residence time (MRT) of epiDX (PlCl: 75.0 l/h, range: 35.6-133.4 l/h; MRT: 31.6 h, range: 7.0-41.5 h;) compared to DX (PlCl: 56.8 l/h, range: 24.4-119.5; MRT: 45.6 h, range: 26.0-83.1 h). No statistically significant differences could be detected for the volume of distribution at steady state (Vss) (epiDX, 31.8 l/kg; DX, 33.3 l/kg). Metabolites common to both compounds were detected in plasma: the 13-dihydro derivatives doxorubicinol (DXol) and epirubicinol (epiDXol), together with minor amounts of four aglycones (7-deoxy adriamycinone, adriamycinone, 7-deoxy 13-dihydro adriamycinone, and 13-dihydro adriamycinone). Following epiDX administration, two additional major metabolites were detected: the glucuronic acid conjugates of epiDX (4'-O-beta-D-glucuronyl-4'-epiDX) and epiDXol (4'-O-beta-D-glucuronyl 13-dihydro-4'-epiDX). This additional detoxication route appears to account for the more efficient and faster elimination of epiDX than of DX. In the urine collected in the 6 days after treatment, 12.2% of the DX and 11.9% of the epiDX dose was excreted as unchanged drug and fluorescent metabolites. A comparable renal clearance was calculated for DX (4.7 l/h, range 1.4-7.0 l/h) and epiDX (4.4 l/h, range 1.7-7.0 l/h). One patient with hepatic metastases and abnormal bilirubin serum level had percutaneous biliary drainage because of extrahepatic obstruction. The elimination of both drugs was significantly impaired in this patient; nevertheless, elimination of epiDX was still more efficient and faster than that of DX (PlCl: 35.6 vs. 24.4 l/h; MRT: 39.0 vs. 83.1 h; t/2 gamma: 47 vs. 74 h). This patient's biliary excretion accounted for 35.4% of the epiDX dose and 18.2% of the DX dose.