Doxorubicin serum elimination kinetics were measured by HPLC in three different patient groups. A dose of (a) 30 mg/m2; (b) 50 mg/m2, and (c) 4 x 15 mg/m2 every 10 h was administered by bolus injection to (a) 10, (b) 6, and (c) 8 patients. The results obtained provided strong evidence for a nonlinear dependence of doxorubicin serum elimination on the dose and administration schedule used. Comparing the 15 and 30 mg/m2 dose there was no significant increase in early drug levels but a marked increase in terminal half-life. At doses higher than 30 mg/m2, however, there was a steep increase in early drug levels, too. Moreover a marked cumulation of the anthracycline in the central compartment following short-term (4 x 15 mg/m2 every 10 h) consecutive administration was found. To obtain an optimal concentration x time product by single bolus injection a dose equal or higher than 30 mg/m2 should be used. However, in this dose range a steep dose-dependent rise in early drug levels is to be expected. As early high serum levels correlate with congestive heart failure, administration schedules reaching effective concentration x time products without high peak levels such as continuous infusion or consecutive administration of low doses seem to be necessary.