Brain EGC, Bachelot T, Serin D, Kirscher S, Graic Y, Eymard J, Extra J, Combe M, Fourme E, Noguès C, Rouëssé J, RAPP-01 Trial Investigators FT. Life-Threatening Sepsis Associated With Adjuvant Doxorubicin Plus Docetaxel for Intermediate-Risk Breast Cancer. JAMA. 2005;293(19):2367-2371. doi:10.1001/jama.293.19.2367
Author Affiliations: Departments of Medical
Oncology (Drs Brain and Rouesse) and Biostatistics (Drs Fourme and Nogues),
René Huguenin Cancer Centre, Saint-Cloud, France; Department of Medical
Oncology, Léon Bérard Cancer Centre, Lyon, France (Dr Bachelot);
Sainte Catherine Institute Breast Clinic, Avignon, France (Drs Serin and Kirscher);
Department of Radiotherapy, Henri Becquerel Cancer Centre, Rouen, France (Dr
Graic); Department of Medical Oncology, Jean Godinot Cancer Centre, Reims,
France (Dr Eymard); Department of Medical Oncology, Institut Curie, Paris,
France (Dr Extra); and Department of Medicine, Centre Hospitalier du Mans,
Le Mans, France (Dr Combe).
Context Adjuvant chemotherapy with new cytotoxic agents for breast cancer must
be properly assessed for toxicity.
Objective To describe adverse events associated with adjuvant chemotherapy for
breast cancer, which led to premature termination of a clinical trial.
Design, Setting, and Patients We conducted a prospective randomized multicenter study (Reposant sur
des Arguments Pronostiques et Prédictifs [RAPP]-01) to compare the
effectiveness of 2 chemotherapy regimens. Patients (women aged 18-70 years)
had primary unilateral breast cancer and either a moderate number of positive
axillary lymph nodes (≤3) or no positive axillary lymph nodes (N0), but
were at a high risk of relapse. Patients were treated at 11 French cancer
referral centers from June 1999 through January 2003. Primary prophylaxis
for febrile neutropenia was not recommended in the study protocol.
Interventions Doxorubicin, 50 mg/m2, plus docetaxel, 75 mg/m2,
or doxorubicin, 60 mg/m2, plus cyclophosphamide, 600 mg/m2, given postoperatively for 4 courses.
Main Outcome Measures The main end point was the disease-free survival rate at 5 years, as
estimated using the Kaplan-Meier product limit method. Secondary end points
included safety, which is the focus of this article, and overall survival.
Results A total of 627 women were enrolled. Median follow-up is currently too
short (24 months) to analyze the primary end point. The trial was terminated
prematurely when 2 deaths related to drug toxicity and 1 case of perforative
peritonitis occurred among patients with febrile neutropenia, all in the doxorubicin-docetaxel
group. The incidence of febrile neutropenia was significantly higher with
the doxorubicin-docetaxel regimen (40.8%) than with the doxorubicin-cyclophosphamide
regimen (7.1%) (P<.001).
Conclusions A high risk of life-threatening complications associated with the doxorubicin-docetaxel
regimen was found in this open-label controlled trial. The doxorubicin-docetaxel
combination should not be considered as an alternative to the doxorubicin-cyclophosphamide
regimen outside carefully designed studies that include primary prophylaxis
for febrile neutropenia.
Anthracyclines are among the most widely used agents for breast cancer
treatment, followed by taxanes, which were gradually introduced during the
last decade. Combinations of these drug classes have proven superior to anthracyclines
alone in advanced or metastatic breast cancer,1,2 while
taxanes are still under evaluation in early breast cancer.3 Uncertainties
regarding the optimal schedule of administration in combination with anthracyclines,
as well as safety and cost issues, are fueling a debate on whether the use
of taxanes is justified outside of clinical trials.4 Febrile
neutropenia is a common adverse effect of chemotherapy, the severity of which
depends both on the types and doses of drugs administered and on individual
Several groups have tested combinations of docetaxel with anthracyclines
in adjuvant and neoadjuvant settings, and the data were recently reviewed.3 Higher rates of febrile neutropenia are reported with
these combinations, including various docetaxel schedules, than with more
traditional regimens incorporating only anthracyclines. In the Breast Cancer
International Research Group (BCIRG) 001 trial, a combination of docetaxel,
75 mg/m2, with doxorubicin and cyclophosphamide caused febrile
neutropenia in 23.9% of patients, compared with 2.4% of patients treated with
5-fluorouracil, doxorubicin, and cyclophosphamide.6 In
the National Surgical Adjuvant Breast and Bowel Project (NSABP) B-27 trial,
21.2% of patients developed febrile neutropenia when docetaxel was administered
sequentially at 100 mg/m2, following anthracyclines and cyclophosphamide,
compared with 7.3% of patients when docetaxel was not included in the treatment.7 These rates are lower than the 40% threshold quoted
in international guidelines warranting the use of granulocyte colony-stimulating
factor (G-CSF) prophylaxis.8,9
Here we report the occurrence of 3 serious adverse events, all associated
with febrile neutropenia and resulting in 2 deaths, during a French multicenter
trial of adjuvant chemotherapy comparing anthracyclines plus cyclophosphamide
vs anthracyclines plus docetaxel in women with early stage breast cancer (Reposant
sur des Arguments Pronostiques et Prédictifs [RAPP]-01 trial). These
events led to premature termination of the trial.
Eligible patients were women aged between 18 and 70 years recruited
in 11 French cancer referral centers who had unilateral operable breast cancer
with clear surgical margins and axillary node clearance, and who were classified
as having high-risk node-negative (N0) or limited node-positive (N+ ≤3)
disease. Written informed consent was required. High-risk N0 status was defined
according to the Saint-Gallen consensus statement.10 Central
randomization was performed by fax or telephone in the Biostatistics Department
of René Huguenin Cancer Center (Saint-Cloud, France), which guaranteed
allocation concealment. It was stratified according to the participating center,
node status (N0 vs N+), and proliferation (Ki67 antigen, <25% vs ≥25%),
using a computerized random-number generator.
The patients were randomly assigned to receive 4 postoperative cycles
of doxorubicin, 50 mg/m2, plus docetaxel, 75 mg/m2 (recommended
dose determined in a phase 1 study11) or 4
postoperative cycles of doxorubicin, 60 mg/m2, plus cyclophosphamide,
600 mg/m2. The latter regimen is a well-established reference schedule
of anthracycline-based chemotherapy. In keeping with international guidelines,8,9 chemotherapy was delivered every 3
weeks without primary G-CSF prophylaxis. Use of G-CSF was recommended only
for grade 3 or 4 febrile neutropenia with a temperature exceeding 38°C
and requiring oral or intravenous antibiotics (National Cancer Institute Common
The primary end point was the disease-free survival rate at 5 years.
To be able to detect a difference in 5-year disease-free survival of 10% (70%
to 80%) with a 1-sided type I error of 5% and a power of 90%, we planned to
recruit 700 patients (350 per group) over 36 months. There were no stopping
rules and no scheduled interim analysis. Toxicity frequencies were compared
by using the χ2 test. P<.05 was
considered significant. The steering committee was in charge of study supervision,
monitoring accrual, treatment compliance, and safety. The protocol was approved
by the institutional review board of each participating center and by the
study ethics review committee.
Enrollments started in June 1999. In March 2000, after inclusion of
45 patients in the doxorubicin-docetaxel group, a 49-year-old patient with
N0 disease and no history of functional gastrointestinal disease developed
latent abdominal pain on day 7 after a first cycle of doxorubicin-docetaxel,
followed within 48 hours by intestinal obstruction, febrile neutropenia, and
suspected mesenteric infarction resulting in death. Autopsy was not performed.
The steering committee did not consider this serious adverse event as specifically
attributable to the taxane-based study regimen and decided that enrollments
In January 2001, after inclusion of 150 patients in the doxorubicin-docetaxel
group, a second case of febrile neutropenia with gastrointestinal disorders
occurred in a 54-year-old woman with N+ disease and no history of abdominal
complaints. She developed perforative peritonitis and septic shock 6 days
after her fourth cycle of doxorubicin-docetaxel. She recovered after extensive
intestinal surgery, which showed diverticulosis, and spent 3 weeks in intensive
care. Enrollments were adjourned pending a survey of similar serious adverse
events in other trials of docetaxel (ongoing or closed to inclusions) and
in the Aventis database. The survey failed to identify factors explaining
the higher incidence of these events in our trial. The written information
and consent forms were modified to mention a potential risk of gastrointestinal
disorders and the ethics committee, independent data monitoring committee,
and the French Medicines Agency authorized enrollments to proceed in July
In January 2003, after enrollment of 311 patients in the doxorubicin-docetaxel
group, a third case of febrile neutropenia with gastrointestinal disorders
occurred after a first cycle of doxorubicin-docetaxel in a 39-year-old patient
with N+ disease. She had a history of nonactive mild diverticulosis and colonoscopic
rectal polyp excision. She developed latent abdominal pain on day 6 and died
on day 13 with septic shock and multiorgan failure, Pseudomonas
aeruginosa septicemia, and diarrhea. Autopsy revealed sigmoiditis with
diverticulitis, partial perforation of the mesocolon, and local peritonitis.
The steering committee terminated the study in January 2003, with 627 patients
enrolled, because of the unexpectedly high toxic mortality rate (2/311 or
0.63%). The 18 patients still receiving doxorubicin-docetaxel were switched
The patients’ characteristics were well balanced between the 2
treatment groups (Table 1). In the doxorubicin-docetaxel
group, however, we found a higher rate of febrile neutropenia (40.8% vs 7.1%
per patient, P<.001; 14.0% vs 1.9% per cycle, P<.001), more frequent G-CSF use, more frequent dose
reductions, and fewer delayed treatment cycles (Table 2). Among patients with grade 3 or 4 gastrointestinal toxicity,
there was less nausea and vomiting, more diarrhea, and more mucositis in the
doxorubicin-docetaxel group (Table 3).
The doxorubicin-docetaxel regimen was associated with more amenorrhea than
the doxorubicin-cyclophosphamide regimen (77.4% vs 54.7%, P<.001). No grade 3 or 4 docetaxel-specific adverse events, defined
as severe fluid retention or nail disorders, were observed.
Of the 87 severe adverse events reported to the French Medicines Agency,
72 (82.8%) occurred in the doxorubicin-docetaxel group, of which 87.5% were
associated with febrile neutropenia (Table 4).
Treatment was completed as planned in the 316 patients assigned to the doxorubicin-cyclophosphamide
group, while treatment was discontinued prematurely in 17 patients in the
doxorubicin-docetaxel group (5.5%) because of adverse events (11 patients)
or patient refusal to continue therapy (6 patients).
Although the 3 severe adverse events described herein did not correspond
to typhlitis or neutropenic enterocolitis sometimes observed with use of docetaxel,13,14 1 patient required extensive intestinal
surgery, 2 patients died, and all 3 cases were associated with febrile neutropenia.
Rare cases of toxic death (0.2%-0.3%) have been reported with various
schedules of adjuvant chemotherapy such as doxorubicin-cyclophosphamide and
combination regimens of cyclophosphamide, methotrexate, and 5-fluorouracil
in multicenter trials conducted in the 1980s and early 1990s.15 However,
acute toxic deaths were generally due to cardiac disorders, even in the absence
of anthracycline use, as in the cyclophosphamide, methotrexate, and 5-fluorouracil
group of the NSABP B-23 trial.15 The rate of
toxic death has decreased far below 0.10% in more recent trials, with high-dose
epirubicin-based adjuvant therapy for example.16 We
observed a much higher rate of toxic death (0.63%) with the doxorubicin-docetaxel
regimen. The higher rate of febrile neutropenia observed with doxorubicin-docetaxel
than with doxorubicin-cyclophosphamide in our trial (40.8% vs 7.1%) may have
induced severe immunosuppression andcontributed to the high rate of toxic
death, which was 3 times as much as that observed in the NSABP B-27 trial,
in which 3 of 7 deaths were attributable to sequential docetaxel immunosuppression
among 1584 patients (0.19%).7
The Spanish GEICAM Group reported that primary G-CSF prophylaxis reduced
the rate of febrile neutropenia induced by adjuvant doxorubicin-docetaxel-cyclophosphamide
therapy from 23.8% to 3.5% and halved the rate of grade 3 or 4 toxicity (50.4%
vs 20%).17 Likewise, in the BCIRG-004 trial,
prophylactic use of growth factors during the doxorubicin-docetaxel-cyclophosphamide
regimen reduced the rate of febrile neutropenia to 7%.18 The
40% incidence of febrile neutropenia in our study is alarmingly high and indicates
an unacceptable toxicity profile. It warrants the use of primary G-CSF prophylaxis,
in accordance with international guidelines.8,9 Prophylactic
antibiotic therapy is unable to reduce the rate of febrile neutropenia induced
by regimens like doxorubicin-docetaxel below the rate (<10%) generally
observed with traditional regimens. For instance, in the doxorubicin-docetaxel-cyclophosphamide
group of the BCIRG-001 adjuvant trial, 24.7% of patients had febrile neutropenia
despite ciprofloxacin prophylaxis.6
Of note, the second and third serious adverse events described herein
occurred in patients with either occult or previously uncomplicated diverticulosis.
The prevalence of diverticulosis increases with age from less than 10% in
persons younger than 40 years to 50% to 66% in those 80 years and older. There
is no clear difference in prevalence between the general population and patients
with breast cancer, but the precise frequency is difficult to measure because
most patients are asymptomatic or present with nonspecific abdominal symptoms
such as constipation and abdominal pain.19 Although
our trial suggests a possible link between preexisting diverticulosis and
the onset of these life-threatening complications, no such relation was found
in large series of similar cases studied in a single institution.14 In our trial, randomization should, in principle,
have led to a similar prevalence of diverticulosis in the 2 groups. Because
routine colonoscopic screening is not feasible in this setting, patients selected
to receive regimens like doxorubicin-docetaxel should have a thorough history-taking,
and general practitioners must be trained to rapidly diagnose these events,
bearing in mind that signs and symptoms of diverticulitis are often more subtle
in immunosuppressed patients than in immunocompetent persons.19 Patients
with compatible symptoms must be referred to a specialist for vigorous supportive
Although encouraging results in similar adjuvant trials6,20,21 support
the routine use of taxanes as a component of standard therapy for node-positive
breast cancer patients, the possibility that more intensive anthracycline-containing
regimens might produce similar results cannot be excluded.22 High-risk
node-negative patients can have a poorer prognosis than patients with limited
node involvement, and one might be also tempted to use taxane-based regimens
for them too, despite the lack of recommendations based on phase 3 clinical
In conclusion, this study shows that the doxorubicin-docetaxel combination
is associated with an increased risk of severe sepsis and life-threatening
complications. Clinicians should be aware of the potential toxicity of the
doxorubicin-docetaxel regimen and consider the preventive use of G-CSF and/or
antibiotics (neither of which was recommended at the time of our trial) in
both the adjuvant and metastatic settings. At this time the doxorubicin-docetaxel
regimen should not be recommended outside of carefully designed clinical trials.
Corresponding Author: Etienne G. C. Brain,
MD, Department of Medical Oncology, René Huguenin Cancer Centre, 35
rue Dailly, 92210 Saint-Cloud, France (firstname.lastname@example.org).
Author Contributions: Dr Brain, as principal
investigator, had full access to all of the data in the study and takes responsibility
for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Brain, Extra, Combe,
Acquisition of data: Brain, Bachelot, Serin,
Kirscher, Graic, Eymard, Rouëssé.
Analysis and interpretation of data: Brain,
Bachelot, Eymard, Extra, Combe, Noguès, Rouëssé, Fourme.
Drafting of the manuscript: Brain, Serin, Graic,
Extra, Combe, Rouëssé.
Critical revision of the manuscript for important
intellectual content: Brain, Bachelot, Serin, Kirscher, Eymard, Extra,
Combe, Noguès, Rouëssé, Fourme.
Statistical analysis: Brain, Extra, Combe,
Noguès, Rouëssé, Fourme.
Obtained funding: Brain, Combe, Rouëssé.
Administrative, technical, or material support:
Study supervision: Brain, Serin, Extra, Combe,
Financial Disclosures: None reported.
Funding/Support: René Huguenin Cancer
Centre was the sponsor of the RAPP-01 trial. This work was also supported
in part by Aventis–Oncology France and the Ligue Régionale Contre
le Cancer du Département des Yvelines (78). Aventis–Oncology
France supplied the docetaxel.
Role of the Sponsor: Staff in the departments
of medical oncology, radiotherapy biology, pathology, and biostatistics of
the René Huguenin Cancer Centre designed and conducted the study, including
collection and analysis of the data, interpretation of the results, and preparation
of the manuscript. Aventis–Oncology France had no role in the preparation
of the manuscript.
Previous Presentation: Presented as a poster
at the 40th annual meeting of the American Society of Clinical Oncology; June
5-8, 2004; New Orleans, La.
Acknowledgment: We thank the investigators
of the RAPP-01 Trial (institutions: René Huguenin Cancer Centre [Saint-Cloud],
Henri Becquerel Cancer Centre [Rouen], Léon Bérard Cancer Centre
[Lyon], Jean Godinot Cancer Centre [Reims], Institut Curie [Paris], Sainte
Catherine Institute Breast Clinic [Avignon], Centre Hospitalier du Mans [Le
Mans], Paul Papin Cancer Centre [Angers], Claude Bernard Clinic [Metz], Edouard
Herriot University Hospital [Lyon], Civil and University Hospitals of Strasbourg
[Strasbourg]) and the members of the independent data monitoring committee
who helped in the decision to submit the manuscript for publication. We thank
I. Stalain-Turbiez, MSc, and M. L. Manche-Thevenot, MSc, for editorial assistance
and monitoring of the data. We thank C. Hill, PhD, for her assistance in the
revision of the text. We express sincere gratitude to all women who participated
in the study.