OBJECTIVE To determine the maximum-tolerable dose (MTD) of paclitaxel given as a 3-hour intravenous (IV) infusion that could be used in conjunction with doxorubicin and dexrazoxane, and to determine the effect of dexrazoxane on the pharmacokinetics of paclitaxel and doxorubicin. METHODS Twenty-five patients with advanced breast cancer received dexrazoxane (600 mg/m2 by IV infusion over 15 minutes), followed 15 minutes later by doxorubicin (60 mg/m2 IV), followed 15 minutes later by paclitaxel (150 or 175 mg/m2 by IV infusion over 3 hours) in cohorts of three to six patients using a standard phase I design without (group A) and with (group B) granulocyte colony-stimulating factor (G-CSF). Treatment continued until there was a substantial decrease in the left ventricular ejection fraction (LVEF), congestive heart failure, progressive disease, or physician discretion to discontinue. RESULTS The MTD of paclitaxel was 150 mg/m2, and adjunctive therapy with G-CSF was required to prevent febrile neutropenia. Dexrazoxane had no significant effect on the pharmacokinetics of paclitaxel or doxorubicin. After a median cumulative doxorubicin dose of 360 mg/m2 (range, 60 to 870 mg/m2), no patient developed congestive heart failure or had a decrease in LVEF below normal. An objective response occurred in all five patients with locally advanced breast cancer and in eight of 20 patients (40%; 95% confidence interval, 19% to 61%) with metastatic breast cancer. CONCLUSIONS When combined with doxorubicin (60 mg/m2) and dexrazoxane (600 mg/m2), paclitaxel given as a 3-hour infusion had an MTD of 150 mg/m2, and G-CSF was required to prevent febrile neutropenia. Dexrazoxane had no effect on the pharmacokinetics of paclitaxel or doxorubicin. No patient in this trial had a decrease in the LVEF below normal, compared with about 20% to 50% of patients treated with doxorubicin and paclitaxel without dexrazoxane in other trials.