The integration of paclitaxel into chemotherapy regimens with anthracyclines offers a new opportunity for devising effective therapy for patients with breast cancer. High response rates have been obtained by combining epirubicin or doxorubicin with paclitaxel. The pharmacokinetic analysis of paclitaxel and anthracyclines, as well as the identification of relationships with their pharmacodynamics, represents a rational approach for treatment optimisation. A schedule-dependent interaction between paclitaxel and anthracyclines has been demonstrated in clinical pharmacokinetic studies. In patients given paclitaxel 125 to 200 mg/m2 as 3- to 24-hour infusions in combination with doxorubicin 48 to 60 mg/m2 as a 48-hour infusion or intravenous bolus, the peak plasma drug concentration (Cmax) of doxorubicin increased significantly and drug clearance was reduced in the sequence paclitaxel-->doxorubicin as compared with doxorubicin-->paclitaxel. The schedule paclitaxel-->doxorubicin was more toxic as compared with doxorubicin-->paclitaxel, and an incidence of 18 to 20% of congestive heart failure was observed in patients with breast cancer given doxorubicin 60 mg/m2 followed by paclitaxel 125 to 200 mg/m2. Likewise, patients given epirubicin 90 mg/m2 had a sudden rebound of epirubicinol plasma concentrations shortly after the start of infusion of paclitaxel 200 mg/m2, with a significant increase in the area under the concentration-time curve (AUC) of epirubicinol as compared with epirubicin alone (1.27 +/- 0.2 vs 0.61 +/- 0.1 mumol/L.h). Moreover, the severity of the myelosuppression induced by paclitaxel, as defined by a sigmoid maximum effect (Emax) relationship between the decrease in neutrophil count and the duration of drug plasma concentrations above the threshold value of 0.1 mumol/L, was significantly enhanced by epirubicin. Finally, chemotherapy with paclitaxel and anthracyclines may be improved by designing pharmacologically guided regimens in order to control the extent of pharmacokinetic interaction and reduce the risk of severe toxicity while maintaining the therapeutic efficacy of the combination. Future protocols should explore the activity of a prolonged paclitaxel infusion in association with an anthracycline separated from the taxane by a washout time interval in order to minimise the inhibitory effects exerted by paclitaxel on P-glycoprotein-mediated biliary clearance of anthracyclines, the most likely cause of pharmacokinetic interaction.