In a pharmacokinetics study, six patients were treated i.v. with epirubicin (EPI) at the two dose levels of 60 and 120 mg/m2, whereas a further six patients were treated at 75 and 150 mg/m2. Both groups were studied according to a balanced cross-over design; the aim of the study was to assess the pharmacokinetic linearity of epirubicin given at high doses. Both the absolute goodness of fit and the Akaike Information Criterion (AIC) point to a linear, tricompartmental open model as the choice framework for discussing EPI plasma disposition after 16/24 administrations, independent of the delivered dose. After 8 treatments, the minimal AIC value corresponded to a nonlinear tissue-binding model. However, even in these cases, second-order effects were present only during the early minutes following treatment. In a model-independent framework, mean EPI plasma clearance was identical at the two dose levels of 60 and 120 mg/m2 (65.4 +/- 8.0 vs 65.3 +/- 13.4 l/h, P = 0.92). Both the mean residence time (MRT) and the volume of distribution at steady-state (VSS) were similar as well (MRT: 22.6 +/- 2.9 vs 24.2 +/- 3.7 h; P = 0.46; VSS: 21.3 +/- 1.5 vs 22.6 +/- 6.5 l/kg, P = 0.46). No statistically significant difference could be found in mean statistical-moment-theory parameters determined after 75- and 150-mg/m2 EPI doses (plasma clearance, PlCl: 83.4 +/- 13.5 vs 68.5 +/- 12.8 l/h, P = 0.12; MRT: 22.6 +/- 4.8 vs 21.9 +/- 3.9 h, P = 0.60; VSS: 26.7 +/- 10.5 vs 21.2 +/- 7.0 l/kg, P = 0.17). Analysis of variance also failed to reveal any significant correlation between dose and plasma clearance. However, when data relative to single patients were examined, a trend toward nonlinear drug distribution as well as a consequent increase in peripheral bioavailability could be observed in 4/6 patients of the 75-mg/m2 vs the 150-mg/m2 group.(ABSTRACT TRUNCATED AT 400 WORDS)