Based on preclinical data, we designed a phase I/II clinical trial to determine the efficacy and toxicity of doxorubicin followed by paclitaxel (Taxol; Bristol-Myers Squibb Company, Princeton, NJ) in the treatment of patients with advanced breast cancer (either untreated or relapsed after adjuvant therapy). In the phase I study, 19 enrolled patients received bolus doxorubicin (50 mg/m2) and, after a 16-hour interval, escalating doses of paclitaxel (from 130 to 250 mg/m2 in 30-mg/ m2 increments) by 3-hour infusion every 3 weeks for a maximum of eight cycles. Paclitaxel doses were increased if the maximum tolerated dose (MTD; defined by dose-limiting toxicities) had not been reached. Analysis of the 128 cycles assessable for toxicity demonstrated neutropenia (< 500/microL) in 20% of cycles with no significant clinical events. No relevant clinical cardiotoxicity was observed. Other toxicities included mild peripheral neuropathy and mild myalgia/arthralgia (In 37.5% and 30.4% of cycles, respectively). The maximum tolerated paclitaxel dose was not reached at the 250 mg/m2 dose level. In the second phase, 13 patients were treated with fixed doses of both drugs (doxorubicin 50 mg/m2 and paclitaxel 220 mg/m2, the dose level immediately preceding the highest paclitaxel dose used in phase I). Grade 4 neutropenia occurred in 36 of the 87 cycles but was complicated by fever in only eight cycles (9%); three patients needed granulocyte colony-stimulating factor. Peripheral neuropathy (grades 1 and 2 in 41.3% and 5.7% of cycles, respectively) and a myalgic syndrome (grades 1 and 2 in 24.1% and 17.2% of cycles, respectively) were observed. No significant clinical cardiotoxicity was observed in 12 of the 13 patients. One patient experienced a decrease in left ventricular ejection fraction (from 60% to 43%) at a cumulative doxorubicin dose of 400 mg/m2. Antitumor efficacy was evaluated in both phase I and phase II. Overall clinical responses included 10 complete (31.3%) and 15 partial (46.9%) responses, for an objective response rate of 78.1%. Six patients (18.8%) had stable disease. The median durations of objective and complete response were 9 and 7 months, respectively. The 78.9% objective response rate in the phase I trial (31.6% complete and 47.3% partial responses) suggests a dose response relationship: at paclitaxel dose > or = 190 mg/m2, all patients had an objective response (six complete and nine partial responses). These results confirm that doxorubicin followed by paclitaxel is active and should be tested as adjuvant treatment and in patients treated previously with anthracyclines.