A Phase II Study of Doxorubicin/Paclitaxel Plus G-CSF for Metastatic Breast Cancer

April 1, 1997

This phase II trial was conducted to evaluate the percentage of objective responses and the toxicity profile of combination doxorubicin (Adriamycin) and paclitaxel (Taxol) with granulocyte colony-stimulating factor as first-line

ABSTRACT: This phase II trial was conducted toevaluate the percentage of objective responses and the toxicity profileof combination doxorubicin (Adriamycin) and paclitaxel (Taxol) with granulocytecolony-stimulating factor as first-line therapy for patients with metastaticbreast cancer (MBC) not previously exposed to anthracycline-containingregimens. Patients with measurable, visceral-dominant MBC and a performancestatus of 0 to 2 were included in the study. Doxorubicin 60 mg/m²was administered as a short intravenous infusion, followed by paclitaxel250 mg/m² as a 3-hour intravenous infusion on day 1. Granulocyte colony-stimulatingfactor 5 µg/kg/d was given prophylactically as a subcutaneous injectionfrom day 2 until granulocyte recovery to 1,500/mm³ or more. Treatmentwas repeated every 21 days for a maximum of six courses. Dose reductions(to doxorubicin 50 mg/m² and paclitaxel 175 mg/m²) and/or treatmentdelay were applied in case of severe toxicity. All 25 women who entered were evaluable for response and toxicity.The main grade 3/4 toxicities observed were leukopenia, thrombocytopenia,and mucositis. Alopecia occurred in all patients. No clinically relevantcardiovascular toxicity was observed. Severe myelosuppression and/or mucositisnecessitated dose reductions at courses 2 or 3 in all but one patient.The complete response rate was 28%, and the partial response ratewas 52% for an overall objective response rate of 80%. Median progression-freesurvival for complete responders was 11 months (range, 3 to 24 months),while the progression-free survival was 7+ months (range 2 to 14+ months)for partial responders and 5 months (range, 3 to 9 months) for nonresponders.This combination produces a high objective response rate in women withMBC, but dose reductions were necessary in almost all cases. Toxicity wasmanageable after dose reduction, allowing patients to be re-treated fortwo to six courses without life-threatening toxicity or toxic deaths. Unfortunately,the duration of response was limited even among complete responders. Furthertrials of this combination in patients with MBC should explore improvementsin this study regimen. [ONCOLOGY11(Suppl):24-29 1997]


Metastatic breast cancer remains incurable with currently availabletherapeutic strategies.[1] Doxorubicin (Adriamycin), a DNA intercalatorthat inhibits topoisomerase II,[2] is considered one of the most activedrugs for single-agent treatment of this disease, producing objective responsesin about 30% to 40% of patients.[3] Combination regimens including doxorubicinmay induce objective responses in 40% to 70% of patients with metastatic breast cancer,although the percentage of complete remissions is only about 5% to 15%,and metastatic breast cancer relapses in virtually all cases within a medianfollow-up time of 6 to 24 months.[4]

Paclitaxel (Taxol), a novel tubulin-interacting agent that promotesthe formation and stabilization of microtubules, has been shown to produceobjective tumor responses in about 20% to 60% of patients with metastaticbreast cancer.[5] Notably, a response rate of about 20% to 30% has beenobserved in heavily pretreated patients.[6] This level of objective responsecompares favorably with that observed with most cytotoxic agents previouslytested in phase II trials against this disease.[1,4]

The combination of doxorubicin and paclitaxel exhibits at least partialsynergism against human breast cancer cell lines in vitro.[7] That observation,added to the comparable levels of antitumor effects of the two agents inpatients with metastatic breast cancer, and the fact that the agents actthrough separate and distinct intracellular mechanisms, prompted the designof a study to evaluate the combined effect of doxorubicin/paclitaxel inwomen with metastatic breast cancer.


Over the last few years, various investigators have studied the single-agentactivity of paclitaxel at different doses and infusion times.[8] In a pilotevaluation conducted at our institution, we administered paclitaxel 250mg/m², given as a 3-hour intravenous (IV) infusion every 3 weeks,to a group of 20 heavily pretreated patients with visceral metastatic breastcancer that had progressed after therapy with a doxorubicin-containingregimen. An objective response rate of 16% was observed and short-lastingreversible neutropenia was the main dose-limiting toxicity encountered.Considering the tolerability of the study regimen in this group of patients,we decided to evaluate the merits of bolus doxorubicin combined with a3-hour IV infusion of paclitaxel as first-line therapy in patients withmetastatic breast cancer.[9] Granulocyte colony-stimulating factor (G-CSF)was added to prevent life-threatening neutropenia, and only women withvisceral-dominant disease were included.


This open-label, nonrandomized phase II study was conducted by the South-AmericanOffice for Anticancer Drug Development in Porto Alegre, Brazil. The trialwas conducted according to guidelines of the local ethics committee andgood clinical research practice. The study was designed to accrue 25 consecutiveand evaluable patients from different institutions. The trial was dividedinto two stages to allow its early discontinuation in the event of severelife-threatening toxicity or lack of efficacy; at least three objectiveresponses among the first 10 patients were required to continue the study.[10]

Patient Selection

Patients with histologically proven metastatic breast cancer, a lifeexpectancy of at least 3 months, and a World Health Organization performancestatus of 0 to 2 were eligible for study entry. Prior adjuvant chemotherapy(excluding anthracyclines) and/or hormonal therapy were allowed, but nocytotoxic therapy for the management of advanced disease was permitted.Radiotherapy for locoregional management of disease and/or palliation ofbone metastases was allowed, as long as patients who had received radiotherapyhad completed treatment at least 3 weeks prior to study entry and all acutetoxic effects of therapy had resolved.

Study subjects had bidimensionally measurable lesions in areas thathad not been irradiated previously. Only patients with predominantly visceraldisease were included in the trial. Patients who had only skin, soft tissue,or lymph node (including supraclavicular) involvement were not eligiblefor the study; patients with brain metastases also were excluded.

Additional eligibility requirements included adequate bone marrow (whiteblood cell count 3,500/mm³or greater, absolute granulocyte count 2,000/mm³orgreater, and a platelet count 100,000/mm³or greater). Patients alsowere required to have adequate hepatic (serum bilirubin level 1.5 mg/dLorgreater) and kidney (serum creatinine level 2.0 mg/dL or greater) function.Patients with active infections or severe concurrent medical conditionswere excluded.

Pretreatment and Follow-Up Evaluation

All patients gave a complete history and underwent physical examination,complete blood cell count, biochemistry analysis (including renal and livertests, and serum electrolytes), urinalysis, chest x-ray, and an electrocardiogram.During drug treatment, patients were re-examined and laboratory tests wererepeated on a weekly basis. Tumor measurements were taken and toxicityassessed before each course of therapy. Additional imaging studies, suchas computed tomography scan or magnetic resonance imaging were repeatedevery two courses to evaluate marker lesions. Whenever possible, the leftventricular ejection fraction was measured by ultrasound or radioisotopictechniques at baseline and after every two treatment courses.

Treatment Plan

Doxorubicin was supplied by Pharmacia Laboratories, Rio de Janeiro,Brazil; paclitaxel was provided by Bristol-Myers Squibb Laboratories, SãoPaulo, Brazil; and G-CSF was supplied by Roche Laboratories, SãoPaulo, Brazil. Doxorubicin 60 mg/m² was administered by IV bolus onday 1, followed immediately by paclitaxel 250 mg/m², administeredas a 3-hour IV infusion. Premedication to avoid allergic reactions to paclitaxelconsisted of dexamethasone 20 mg IV, ranitidine 50 mg IV or cimetidine200 mg IV, and diphenhydramine 100 mg IV or promethazine 50 mg intramuscularly,30 minutes prior to paclitaxel administration. As part of the antiemeticregimen, dexamethasone was repeated at 10 mg IV after 4 and 8 hours, andthen daily at 4 mg orally every 8 hours for an additional 3 days. Cimetidine200 mg IV was given after 8 hours on day 1 and maintained for 3 days at300 mg orally every 8 hours to prevent gastrointestinal complications ofdexamethasone. Diphenhydramine 100 mg IV was repeated once after 8 hourson day 1.

Prophylactic G-CSF 5 µg/kg was administered daily as a subcutaneousinjection, starting on day 2 and continuing until the granulocyte countreached 1,500/mm³. Treatment was repeated every 3 weeks for a maximumof six cycles, after which patient management was decided by the individual'sphysician. In the case of a granulocyte count lower than 1,500/mm³and/or a platelet count less than 100,000/mm³ on day 21, treatmentwas postponed for 1 week. If counts had returned at least to lower normallimits by day 28, treatment was repeated at the same dose level. If recoverywas documented only after a 2-week delay (ie, by day 35), or if the patientdeveloped severe mucositis and/or neutropenic fever necessitating antibioticsand hospitalization, treatment was repeated at a reduced dose, arbitrarilyset at doxorubicin 50 mg/m² and paclitaxel 175 mg/m². The paclitaxeldose was based on preliminary information obtained from other concomitanttrials.[5,6,14] Given that complete information on the safety of this doselevel was not yet completed during the treatment period for the currentstudy, G-CSF was also maintained according to the study protocol. Patientswhose treatment was interrupted for more than 2 weeks were not re-treatedin the protocol but were managed on the basis of clinical judgment.

Response and Toxicity Evaluation

Identification of objective response was based on the response definitionsestablished by the World Health Organization. Progression-free survivalwas calculated as the time from first documentation of objective response(complete [CR] or partial [PR]) to first documentation of disease progression.Toxicity evaluation was based on the National Cancer Institute Common ToxicityCriteria.

Patient Characteristics

Between September 1992 and May 1995, 25 patients were accrued to thisstudy. The trial was temporarily interrupted twice due to a shortage inthe paclitaxel supply that resulted in a total of 11 months without patientaccrual. However, this fact did not seem to influence patient selectionor introduce any detectable bias into the protocol study.

The main characteristics of this group of patients are listed in Table1. All 25 women had visceral-dominant disease; World Health Organizationmedian performance status of 1 (range, 1 to 2), and a median age of 49years (range, 29 to 65 years), with only three patients older than 60 years.All patients were initially treated by radical modified mastectomy or conservativesurgery, followed by axillary lymph node dissection plus radiotherapy.No patient had undergone prior hormonal, immunologic, or cytotoxic therapyfor advanced disease. Adjuvant cyclophosphamide-methotrexate-5-fluorouraciltherapy had been given to 19 of the 25 patients.

Toxicity Profile and Dose Administrations

The main toxic effects observed in the trial were neutropenia, thrombocytopenia,mucositis, and total alopecia. Less common mild to moderate side effectswere myalgia/arthralgia, nausea and vomiting, peripheral neuropathy, isolatedcases of reversible mild hypotension, and short-lasting and reversible skin papuloerythematous reactionduring paclitaxel infusion. These findings are summarized in Table2.

In spite of prophylactic G-CSF administration, and with the exceptionof one patient whose disease progressed after the first course of therapy,all patients required dose reduction after the second or third treatmentcourse. In six patients, severe mucositis plus neutropenic fever necessitatedsupportive care and/or hospitalization. In all other patients, dose reductionswere based on the duration of neutropenia with or without fever, clinicalsigns of septicemia, and/or severe mucositis. No toxic deaths were documentedin the study. Dose reduction was needed for one of three patients receivingtwo courses of therapy, and for all patients who received more than twocourses of therapy.

Thus, of 102 courses studied, 27 were delivered at the initial planneddose (doxorubicin 60 mg/m², paclitaxel 250 mg/m²), and 75 wereadministered at the reduced dose (doxorubicin 50 mg/m², paclitaxel175 mg/m²). At the higher dose level, grade 3/4 neutropenia occurredin all courses (grade 4, 93%), and grade 3/4 thrombocytopenia occurredin 67% of courses (grade 3, 63%). Grade 3 mucositis was observed in 37%of courses. In contrast, grade 3/4 neutropenia was documented in 65% ofcourses at the reduced dose level (grade 4, 35%), and thrombocytopeniawas seen in only 16% (grade 4, 4%). Grade 3 mucositis was observed in 3%of courses.

In general, repeated courses of the reduced-dose regimen given withG-CSF were well tolerated, with 17 of 25 patients receiving four or moretreatment courses. It is worth noting that no case of life-threateningneutropenia-related infection or severe mucositis requiring hospitalizationfor IV hydration was observed in this group of patients.

As part of the initial locoregional treatment, 20 of the 25 patientsreceived radiation to the left anterior chest wall, internal mammary nodes,and axilla, but no other detectable risk factor for heart disease was presentin the study population. The median cumulative dose of doxorubicin was220 mg/m² (range, 60 to 320 mg/m²).

Unfortunately, follow-up cardiac toxicity evaluation by echocardiogramand/or multiple gated image acquisition analysis was not performed systematicallyin all patients. In 16 patients, however, at least one left ventricularejection fraction measurement was taken prior to entry into the trial andwas repeated after two to six courses either by multiple gated image acquisitionanalysis scan or echocardiogram. In two cases, reductions of 12% and 14%were observed in relation to baseline values. However, no clinical signsof cardiotoxicity, such as the development of congestive heart failure,cardiac arrhythmia, or angina were seen. Both patients had received priorchest radiation and their cumulative doxorubicin doses were 270 and 320mg/m2. The two patients were removed from the study and continuedtreatment with paclitaxel alone, one for an additional three courses andthe other for two courses.

Efficacy and Progression-Free Survival

Objective responses (CR-PR) were documented in 80% of patients, with28% (seven cases) achieving a complete response (Table3). Of the seven patients who achieved a CR, four were referred toexperimental high-dose chemotherapy programs for consolidation after thefourth course. Unfortunately, disease relapsed in all four patients at3, 7, 11, and 16 months, following the latter procedure. Responding patients(both CR and PR) who were not referred to high-dose chemotherapy programswere treated with paclitaxel alone for additional courses, switched toa different chemotherapy regimen, or maintained on tamoxifen 20 mg/d ormegestrol acetate 160 mg/d.

Progression-free survival for the whole group of patients, as well asfor responders and nonresponders, is depicted in Figure1. Relapses occurred most commonly in visceral sites, while centralnervous system metastasis was the site of initial failure in three cases.After a 24-month follow-up, only one patient (from the PR group) remainedprogression free. This patient received four courses of therapy and thendecided to withdraw from the study for treatment with mitoxantrone forthree courses, after which she was maintained on tamoxifen therapy. Themedian progression-free survival time was 11 months (range, 3 to 24 months)for patients who achieved a CR, 7+ months (range, 2 to 14+ months) forthose with a PR, and 5 months (range, 3 to 9 months) for nonrespondingpatients.


Paclitaxel as first- or second-line chemotherapy has been associatedwith high objective response rates in women with metastatic breast cancer.[5]Paclitaxel also has been shown to produce objective responses in 20% to40% of patients who fail to respond to doxorubicin-containing regimens,[6]suggesting a lack of complete clinical cross-resistance between the twoagents.

Although preliminary data on the combined use of doxorubicin and paclitaxelhave revealed a high percentage of objective responses in patients withmetastatic breast cancer, the potential toxicity of the combination isdisturbing.[11-13] Of particular concern are possible pharmacokinetic interactionsbetween the agents, superimposable dose-limiting toxicities to the bonemarrow, and, more specifically, potential risks of cardiotoxicity.[14-16]

Recently, Gianni and colleagues, from the National Tumor Institute inMilan, Italy, reported on the high efficacy of combination paclitaxel by3-hour infusion plus bolus doxorubicin in women with previously untreatedmetastatic breast cancer.[14] The authors observed a 41% complete responserate and a 94% overall response rate. After a median follow-up of 12 months(range, 3 to 18 months), the median response duration was 11 months (range,2+ to 18+ months) for patients who achieved a complete response and 8 months(range 1+ to 15+ months) for those with a partial response. These resultsare in accordance with our report, considering our overall objective responserate of 80% and our complete response rate of 28%. Further, the medianprogression-free survival in our study was 11 months for those attaininga complete response and 7+ months for those with a partial response.

Our study confirms the high level of antitumor activity of the doxorubicin/paclitaxelcombination in patients with metastatic breast cancer, particularly meaningfulin that our study population consisted totally of patients with visceraldisease. Only about 20% of patients receiving paclitaxel/doxorubicin in the Milan trial had soft tissueinvolvement. The limited number of patients and large confidence intervals,as well as distinct patient characteristics, probably account for the observeddifferences in CR rate (28% vs 41% in the Italian study) between thesetwo studies.

Toxicities of the Regimen

In addition to severe neutropenia and mucositis (World Health Organizationgrade 3/4), which occurred in all except one patient, grade 2 peripheralneuropathy was observed in one third of the cases in the Milan trial. Further,clinically reversible congestive heart failure (New York Heart Associationgrades II to III) developed in six patients (18%), five of whom receiveda cumulative doxorubicin dose of 480 mg/m². A history of high bloodpressure was documented in four cases, a family history of heart failureand irradiation of the left breast in one patient each. Although it hasbeen shown previously that paclitaxel can modify the pharmacokinetics ofdoxorubicin and its main metabolite, doxorubicinol, thereby enhancing theadverse effects of the drug,[15,16] the authors did not see any influenceof drug sequence on the antitumor or toxic effects of the study regimen.[14]

Despite concomitant G-CSF administration, severe neutropenia and, lessoften, mucositis necessitated dose reductions (to doxorubicin 50 mg/m²and paclitaxel 175 mg/m²) in all patients who received two or moretreatment courses. With this modified regimen, patients could be re-treatedat 21-day intervals, without further compromising the dose intensity ofthe two drugs. As G-CSF was given to all patients according to the studyprotocol, its value for preventing life-threatening infections in the specificcontext of dose reduction cannot be evaluated. Based on the experienceof Gianni and colleagues,[14] however, G-CSF might not be necessary forpatients treated at the reduced dose level. This aspect of the study deservesfurther evaluation.

Likewise, the median cumulative dose of doxorubicin given to patientsin our trial was 220 mg/m² (range, 60 to 320 mg/m²), which couldexplain the lack of clinical cardiotoxicity. In fact, our strategy wasto administer a maximum of six cycles of therapy even to responding women.Patients younger than 50 years with good performance status were referredfor high-dose therapy after completing four courses of therapy; other respondingpatients (CR and PR) switched to other chemotherapy regimens or maintenancetherapy with tamoxifen or megestrol acetate after they completed theirsixth course. Notably, most patients, including those who exhibited a slightdecrease in the left ventricular ejection fraction, had irradiation ofthe left chest wall as part of their primary management, but had no otherdetectable risk factor for the development of heart disease.

Interestingly, no major allergic reactions developed in our patientpopulation. This observation is notable because the premedication regimenwe used was slightly different from that used routinely with paclitaxel,consisting of antiallergic agents given only 30 minutes before startingthe infusion. In addition, peripheral neuropathy was not a significantproblem in our study, probably due primarily to the small number of paclitaxelcourses administered. Also, however, these patients had no prior exposureto neurotoxic agents and no known predisposing medical condition, likechronic alcohol abuse, diabetes, or other metabolic disorders. Alopeciawas universal, as would be expected for this drug combination. Other toxicitieswere easily managed and of no clinical importance.


Our trial of combination doxorubicin/paclitaxel given as first-linetreatment for patients with visceral-dominant metastatic breast cancerproduced high objective response rates, with manageable toxicity followingdose reduction. The need for G-CSF support in the reduced-dose regimenremains unclear. Unfortunately, progression-free survival was short, evenin the subgroup of patients who had achieved a complete response. Thisobservation implies that, although the two-drug combination achieved markedtumor cell kill, a significant number of resistant tumor cells remainedviable to produce disease progression. Thus, other active agents, withdifferent mechanisms of action, are still needed to achieve long-lastingresponses in patients with this disease.

This paper was presented in abstract form at the annualmeeting of the American Society of Clinical Oncology, Philadelphia, PA,May, 1996.


1. Henderson IC: Breast Diseases, 2nd ed. pp 604-665.Philadelphia, Lippincott, 1991.

2. McVie JG, Schwartsmann G: Oxford Textbook of Oncology, pp513-523. New York, Oxford University Press, 1995.

3. Schwartsmann G, Pinedo HM: Clinical trials in advanced breast cancer.Eur J Cancer 23:595-597, 1987.

4. Osborne CK: Manual of Oncologic Therapeutics 1989/1990, pp201-211. Philadelphia, Lippincott, 1989.

5. Seidman AD: The emerging role of paclitaxel in breast cancer therapy.Clin Cancer Res 1:247-256, 1995.

6. Seidman AD, Tiersten A, Hudis C, et al: Phase II trial of paclitaxelby 3-hour infusion as initial and salvage chemotherapy for metastatic breastcancer. J Clin Oncol 13:2575-2581, 1995.

7. Koechli OR, Sevin BU, Perras JP, et al: Characteristics of the combinationpaclitaxel plus doxorubicin in breast cancer cell lines analyzed with ATP-cellviability assay. Breast Cancer Res Treat 28:21-27, 1993.

8. Arbuck SA: Paclitaxel: What schedule? What dose? J Clin Oncol12:233-235, 1994.

9. Schwartsmann G, Menke CH, Caleffi M, et al: Phase II trial of Taxol,doxorubicin plus G-CSF in patients with metastatic breast cancer (abstract).Proc Am Soc Clin Oncol 15:126, 1996.

10. Schwartsmann G, Wanders J, Koier I, et al: EORTC New Drug Development Coordinating Committee guidelines for the performanceof phase I and II clinical trials. Eur J Cancer 9:1162-1168, l991.

11. Gianni L, Munzone E, Capri G, et al: Paclitaxel in metastatic breastcancer: A trial of two doses by 3-hour infusion in patients who failedprior therapy with anthracyclines. J Natl Cancer Inst 87:1169-1175,1995.

12. Sledge GW Jr, Robert N, Sparano JA, et al: Paclitaxel/doxorubicincombinations in advanced breast cancer: The Eastern Cooperative OncologyGroup experience. Semin Oncol 21(suppl 8):15-18, 1994.

13. Rowinsky EK, McGuire WP, Guarnieri T, et al: Cardiac disturbancesduring the administration of Taxol. J Clin Oncol 9:1704-1712, 1991.

14. Gianni L, Munzone E, Capri G, et al: Paclitaxel by 3-hour infusionin combination with bolus doxorubicin in women with untreated metastaticbreast cancer: High antitumor efficacy and cardiac effects in a dose-findingand sequence-finding study. J Clin Oncol 13:2688-2699, 1996.

15. Gianni L, Kearns CM, Gianni A, et al: Nonlinear pharmacokineticsand metabolism of paclitaxel and its pharmacokinetic/pharmacodynamic relationshipsin humans. J Clin Oncol 13:180-190, 1995.

16. Berg SL, Cowan KH, Balis FM, et al: Pharmacokinetics of Taxol anddoxorubicin administered alone and in combination by continuous 72-hourinfusion. J Natl Cancer Inst 86:143-145, 1994.