Assessment of Cardiac Dysfunction in a Randomized Trial Comparing Doxorubicin and Cyclophosphamide Followed by Paclitaxel, With or Without T
http://www.100md.com
《临床肿瘤学》
the National Surgical Adjuvant Breast and Bowel Project (NSABP)
NSABP Biostatistical Center, University of Pittsburgh
Allegheny General Hospital
University of Pittsburgh Cancer Institute, Breast Cancer Institute, and Magee Women's Hospital, Pittsburgh, PA
Cancer Research Network, Inc, Plantation, FL
University of Kentucky Markey Cancer Center, Lexington, KY
The University of Texas, M.D. Anderson Cancer Center, Houston, TX
Novartis Pharmaceutical Corporation, East Hanover, NJ
Cancer Therapeutics Branch, National Cancer Institute, Bethesda, MD
Kaiser Permanente Vallejo, Vallejo, CA
Atlanta Cancer Care, Atlanta, GA
Aultman Hospital, Canton, OH.
ABSTRACT
PURPOSE: Trastuzumab is effective in treating human epidermal growth factor receptor 2 (HER2) –positive breast cancer, but it increases frequency of cardiac dysfunction (CD) when used with or after anthracyclines.
PATIENTS AND METHODS: National Surgical Adjuvant Breast and Bowel Project trial B-31 compared doxorubicin and cyclophosphamide (AC) followed by paclitaxel with AC followed by paclitaxel plus 52 weeks of trastuzumab beginning concurrently with paclitaxel in patients with node-positive, HER2-positive breast cancer. Initiation of trastuzumab required normal post-AC left ventricular ejection fraction (LVEF) on multiple-gated acquisition scan. If symptoms suggestive of congestive heart failure (CHF) developed, source documents were blindly reviewed by an independent panel of cardiologists to determine whether criteria were met for a cardiac event (CE), which was defined as New York Heart Association class III or IV CHF or possible/probable cardiac death. Frequencies of CEs were compared between arms.
RESULTS: Among patients with normal post-AC LVEF who began post-AC treatment, five of 814 control patients subsequently had confirmed CEs (four CHFs and one cardiac death) compared with 31 of 850 trastuzumab-treated patients (31 CHFs and no cardiac deaths). The difference in cumulative incidence at 3 years was 3.3% (4.1% for trastuzumab-treated patients minus 0.8% for control patients; 95% CI, 1.7% to 4.9%). Twenty-seven of the 31 patients in the trastuzumab arm have been followed for 6 months after diagnosis of a CE; 26 were asymptomatic at last assessment, and 18 remained on cardiac medication. CHFs were more frequent in older patients and patients with marginal post-AC LVEF. Fourteen percent of patients discontinued trastuzumab because of asymptomatic decreases in LVEF; 4% discontinued trastuzumab because of symptomatic cardiotoxicity.
CONCLUSION: Administering trastuzumab with paclitaxel after AC increases incidence of CHF and lesser CD. Potential cardiotoxicity should be carefully considered when discussing benefits and risks of this therapy.
INTRODUCTION
Trastuzumab (Herceptin; Genentech, South San Francisco, CA) is a humanized monoclonal antibody targeting the extracellular domain of the human epidermal growth factor receptor 2 (HER2) protein. Amplification or overexpression of HER2 is seen in approximately 25% of invasive breast cancers and is associated with unfavorable prognosis.1 Trastuzumab was approved for treatment of HER2-positive metastatic breast cancer in 1998 based on improved outcomes in the pivotal trial that evaluated trastuzumab in combination with chemotherapy. However, an increased incidence of cardiac dysfunction (CD) was observed, particularly when trastuzumab was combined with doxorubicin/cyclophosphamide (AC).2
As part of the process of filing for US Food and Drug Administration approval for use in metastatic breast cancer, an independent Cardiac Review and Evaluation Committee (CREC) was convened to review the cardiotoxicity data from all metastatic disease trials evaluating trastuzumab and established the following criteria for diagnosis of CD: (1) cardiomyopathy characterized by a global decrease in left ventricular ejection fraction (LVEF); (2) signs or symptoms of congestive heart failure (CHF); (3) decline in LVEF of at least 5% to less than 55% with signs or symptoms of CHF or decline in LVEF of at least 10% to less than 55% without signs or symptoms of CHF. Episodes of CD were characterized according to the New York Heart Association (NYHA) functional classification.3 AC alone was associated with an overall rate of CD of 8.1% and a class III or IV rate of 3.7%. Addition of trastuzumab to AC was associated with an increase in overall CD rate to 27.3% and an increase in the rate of class III or IV CD to 16.1%. Administration of trastuzumab with paclitaxel in patients with prior anthracycline exposure resulted in an overall CD rate of 13.2% but a class III/IV rate of only 2.2%. The CREC also reviewed patients after treatment for CD when possible. Of 34 patients who developed CD on AC plus trastuzumab, seven had persistent class III or IV symptoms on follow-up compared with 0 of 10 patients who developed CD on paclitaxel plus trastuzumab.3
The pivotal trial results justified evaluating trastuzumab with chemotherapy in the HER2-positive adjuvant patient population. This required a trial design that would minimize risk and evaluate the incidence, severity, and reversibility of CD. The low rate and apparent reversibility of class III or IV CHF that was observed when trastuzumab was combined with paclitaxel supported the comparison of AC followed by paclitaxel ± trastuzumab.
Therefore, National Surgical Adjuvant Breast and Bowel Project (NSABP) trial B-31 included two stages, with assessment of CD as the primary aim of the first stage. Stage 1 included the first 1,000 patients who received post-AC treatment and were evaluable for cardiac assessment. After favorable review of the stage 1 cardiac data by an independent data monitoring committee (DMC), accrual was extended to the second stage to enroll an additional 1,700 patients. A recent planned joint interim analysis of B-31 and North Central Cancer Treatment Group (NCCTG) N9831 demonstrated a 52% reduction in hazard of first events (recurrence, second primary, or death; hazard ratio [HR] = 0.48; 95% CI, 0.39 to 0.59; P = 2 x 10–12) and a 33% reduction in hazard of mortality (HR = 0.67; 95% CI, 0.48 to 0.93; P = .015), prompting the DMCs of both groups to recommend disclosure.4 We report here the detailed cardiac safety data from B-31 as of that analysis.
PATIENTS AND METHODS
Patient Eligibility and Entry Procedures
Patients must have had histologically node-positive HER2-positive primary breast cancer without evidence of metastatic disease and must have completed primary surgery and axillary dissection. LVEF on multiple-gated acquisition (MUGA) scan had to be the lower limit of normal (LLN). Patients with active cardiac disease, including angina pectoris or cardiac arrhythmia requiring medication, severe conduction abnormalities, clinically significant valvular disease, cardiomegaly, ventricular hypertrophy, or poorly controlled hypertension, were ineligible. Patients with prior myocardial infarction, CHF, or other cardiomyopathy were also excluded. Protocol approval by the local human investigations committee or IRB was required, with an assurance filed with the Department of Health and Human Services. Written informed consent was required.
Protocol Treatment Regimens
Regimens consisted of either doxorubicin (60 mg/m2) and cyclophosphamide (600 mg/m2) every 21 days for four cycles followed by paclitaxel (175 mg/m2) every 3 weeks for four cycles (arm 1) or the same chemotherapy plus weekly trastuzumab starting with the first dose of paclitaxel at a loading dose of 4 mg/kg followed by 2 mg/kg for 51 weeks (arm 2). (Subsequent to May 16, 2003, paclitaxel could be administered weekly at 80 mg/m2 for 12 weeks.) Patients with lumpectomies received breast irradiation after chemotherapy. Postlumpectomy regional and postmastectomy locoregional irradiation were optional. Trastuzumab was continued during radiation. Radiation of internal mammary nodes was prohibited. Women with hormone receptor–positive tumors were to receive tamoxifen for 5 years.
Cardiac Monitoring
Cardiac history forms were to be submitted at entry, every 6 months for the first 5 years, and yearly thereafter. MUGA scans were required in both treatment arms before entry, after AC, and at 6, 9, and 18 months. Additional scans could be performed at investigator discretion.
Initiation of Trastuzumab and Treatment Discontinuation
Trastuzumab was initiated in arm 2 patients if they remained without cardiac symptoms after AC and their LVEF remained LLN and had not decreased by more than 15 percentage points from the pre-entry level. Criteria were established for holding trastuzumab in asymptomatic patients if 6- and 9-month LVEF values did not remain within predefined limits, as listed in Table 1. If a participant's LVEF did not meet criteria for continuing trastuzumab, the drug was held for 4 weeks. If a repeat LVEF did not meet criteria for continuation, further administration was prohibited.
Monitoring and Review of Symptomatic CD
We will use the term CHF to denote NYHA class III or IV CD. Patients with lesser dysfunction are not referred to as having experienced CHF, although that term is sometimes used to refer to lesser degrees of CD in routine clinical practice.
All symptoms suggestive of CHF were reported within 14 days. Source documents were blinded to treatment arm and reviewed by an independent Cardiac Review Panel consisting of three cardiologists who determined by majority vote whether each patient met prospectively defined criteria for CHF. Criteria mirrored the CREC criteria as follows: NYHA class III or IV symptoms with either a decrease from baseline in LVEF of more than 10 percentage points to less than 55% or a decrease of more than 5 percentage points to less than the LLN. Class III is characterized by symptoms of dyspnea with activities such as climbing a flight of stairs, whereas class IV symptoms are present at rest.
Patients for whom potential CD was reported continued to be monitored by MUGA scans and cardiac follow-up forms that summarized ongoing symptoms and cardiac medications.
Formal Interim Analysis of Cardiotoxicity
The primary cardiac safety end point was the difference in incidence of cardiac events (CE) between arms, defined as confirmed class III or IV CHF or possible/probable cardiac death. During study planning, it was judged that at least a 25% reduction in hazard of mortality was not unlikely with this treatment. This translates to an 8% absolute survival benefit at 10 years (70% v 62%). Given this level of expected benefit, there was agreement that a difference in incidence of CE of up to 4% would be acceptable to allow completion of accrual in stage 2. Cardiac interim analyses were scheduled after 200, 600, and 1,000 v patients had been followed for at least 6 months after the first day of treatment with paclitaxel ± trastuzumab. If at any interim analysis the difference in CE occurrence was statistically greater than 4% (one-sided P < .05), accrual was to be suspended. Otherwise, accrual would continue to a target of 2,700 patients to assess efficacy. This interim analysis plan had a 12% probability of terminating accrual if the true difference in event rate was 4% and an 85% probability of terminating accrual if the difference was 7%.
Statistical Methods
This report is based on all patients accrued as of February 15, 2005 and all cardiac follow-up data received by April 21, 2005. Primary comparisons are based on the cohort of patients who completed AC and began post-AC treatment with paclitaxel (arm 1) or trastuzumab plus paclitaxel (arm 2). To ensure comparability, all patients in both arms of this evaluable cohort were required to meet the protocol conditions for initiating trastuzumab that were required of arm 2 patients (ie, patients were without cardiac symptoms after AC and had a post-AC LVEF LLN with a 15 percentage points decline from baseline). Patients in this cohort were considered to be at risk for CE from day 1 of cycle 5 until recurrence, diagnosis of second primary cancer, or last follow-up. Cumulative incidence was estimated nonparametrically.5 The difference in cumulative incidence of CE in evaluable arm 1 and arm 2 patients 3 years after day 1 of cycle 5 was compared with the hypothesized difference of 4% using a one-sided 0.05-level test. The ratio of crude hazards for CE was estimated by fitting a Cox model. Secondary intent-to-treat analyses were performed using the cohort of all patients with cardiac follow-up.
Association of risk factors with CE was tested using univariate log-rank analysis. A permutation method was used to compute P values.6 In exploratory analyses, Cox models were fit to assess the independence of predictors. Changes in LVEF from baseline were tested using the signed rank procedure; comparisons of LVEF between arms were tested using the rank sum test applied to changes from baseline.
RESULTS
Accrual and Patient Characteristics
The trial opened on February 21, 2000. Cardiotoxicity interim analyses were reviewed by the DMC on October 1, 2001, October 11, 2002, and October 10, 2003. In all three analyses, the difference in CE rates was less than 4%, so patient accrual continued.7 By February 15, 2005, 2,043 patients had been enrolled. Figure 1 summarizes their treatment, evaluation, and follow-up. Twenty-two patients refused protocol therapy; AC treatment information is pending for 49 patients; 148 patients did not meet the post-AC cardiac criteria for initiation of trastuzuma and were excluded from the primary cardiac safety analysis, regardless of assigned treatment arm; 36 patients had pending post-AC MUGA scans; and 124 patients did not begin post-AC therapy or had no post-AC follow-up as of this report. The remaining 1,664 patients (81%) are included in the evaluable cohort. The median follow-up of patients in the evaluable cohort is 27 months from the first day of cycle 5. Characteristics of evaluable patients are balanced by treatment arm. Fifty-one percent of patients were 49 years old at entry, 33% were 50 to 59 years old, and 16% were 60 years old. Eighty-four percent of patients were white, 8% were black, and 8% were of other or unknown race/ethnicity. Fifty-seven percent of patients had one to three positive nodes, 29% had four to nine positive nodes, and 14% had 10 positive nodes. Fifty-nine percent of patients had a pathologic tumor size of more than 2.0 cm. Fifty-two percent of patients had estrogen receptor–positive tumors.
Cumulative Incidence of CEs
Among 814 evaluable control patients, five met criteria for a CE (four CHFs and one probable cardiac death). In the arm 2 evaluable cohort, 31 of 850 patients experienced a CE (31 CHFs and no cardiac deaths). Cumulative incidence of CEs in the control arm 3 years after day 1 of cycle 5 was 0.8% (95% CI, 0.3% to 1.9%). Among trastuzumab-treated evaluable patients, the cumulative incidence of CEs 3 years after day 1 of cycle 5 was 4.1% (95% CI, 2.9% to 5.8%; Fig 2). The difference between the arms was 3.3% (95% CI, 1.7% to 4.9%), which was in agreement with the 4% anticipated increase (P > .5). The relative risk of a CE was 5.9 in trastuzumab patients versus control patients (95% CI, 2.3 to 15.3; P < .0001). Three of the 31 cases of CHF in arm 2 occurred more than 1 year after initiation of trastuzumab.
Five additional CEs occurred in patients not included in the evaluable cohort (three patients in arm 1 and two in arm 2; Fig 1). In arm 2, a patient died of CHF after her third cycle of AC. A second patient with multiple cardiac risk factors died of cardiopulmonary collapse after her first dose of AC. Neither patient received paclitaxel or trastuzumab. In arm 1, three patients with post-AC LVEF less than LLN received paclitaxel and subsequently experienced CHF. (These patients were not protocol violations because normal post-AC LVEF was not required to initiate paclitaxel, although it was required in arm 2 patients to receive trastuzumab.) The intent-to-treat 3-year cumulative incidence of CEs in all arm 1 patients was 1.1% (95% CI, 0.6% to 2.3%) compared with 3.9% in the entire arm 2 cohort (95% CI, 2.8% to 5.5%).
Recovery of Cardiac Function After CHF
Among the 31 evaluable patients meeting criteria for CHF who received trastuzumab, none have died for reasons other than metastatic breast cancer. Twenty-seven of the 31 patients have been followed for 6 months after diagnosis of CHF, of whom 26 have been without symptoms of CD for at least 6 months. However, 18 patients continued to receive cardiac medications (Table 2). Twenty-four of the 27 patients have reported LVEF assessments 6 months after the diagnosis of their CHF, and of these 24 patients, 17 had decreased LVEF relative to baseline at their last available MUGA. The most recent LVEF values were between 40% and 69% in all except one patient whose LVEF was 25% (Fig 3B). Of the five confirmed patients with CHF in arm 1, three have died (probable cardiac death, metastatic breast cancer, emphysema). One of the two surviving patients has been followed for 6 months after her diagnosis of CHF and is currently symptomatic and receiving medications (Table 2).
Recovery of Cardiac Function in Patients With Symptoms of Possible CD That Did Not Meet Criteria for CE
In addition to those patients with confirmed CHF, 43 evaluable arm 2 patients reported symptoms of CD that did not meet protocol criteria for a CE. Thirty-nine patients have been followed for 6 months since onset of symptoms, of whom one patient reported symptoms of CD at last follow-up; eight patients continued to receive cardiac medication (Table 2). MUGA results have been reported 6 months after onset of symptoms for 30 of the 39 patients. Current LVEF results range from 45% to 86%, except for one patient whose LVEF was 40% (Fig 3D). Twenty-two of these patients had decreased LVEF relative to baseline at last evaluation, suggesting some decrement in function.
In addition to the five evaluable arm 1 patients with confirmed CEs, eight evaluable arm 1 patients reported symptoms of CD before recurrence that did not meet protocol criteria for a CE. Six patients have been followed for 6 months since onset of symptoms; all patients were asymptomatic and receiving no cardiac medications (Table 2).
Discontinuation of Trastuzumab Therapy in Patients in Arm 2
Trastuzumab treatment information was submitted quarterly. At least one treatment form has been submitted on 825 of 850 evaluable arm 2 patients, and 714 have completed trastuzumab, and all required documentation has been submitted. Trastuzumba was permanently discontinued before completion of 1 year of therapy or recurrence in 197 patients (28%) (Table 3). Trastuzumab was discontinued for cardiac-related reasons (asymptomatic decline in LVEF, symptoms of CD, or other cardiac symptoms) in 133 patients (19%), including 17 (2%) of 825 patients during the first quarter of therapy, 57 (8%) of 757 patients during the second quarter, 44 (7%) of 643 patients during the third quarter, and 15 (3%) of 538 patients during the final quarter.
Of 102 evaluable patients who discontinued trastuzumab because of asymptomatic declines in LVEF, 21 subsequently reported symptoms of possible CD and have been included in the previous cohorts. Of the remainder, LVEF has been reported in 52 patients 6 months after discontinuation. Their nadir and most recent LVEF results are shown in Figures 3E and 3F, where they are contrasted with the nadir and most recent results for trastuzumab-treated patients with confirmed CEs (Figs 3A and 3B) and patients who had symptoms of CD not meeting criteria for a CE (Figs 3C and 3D). Of these 52 patients, the most current LVEFs are less than 45% in three patients (6%) and less than 50% in 13 patients (25%).
Trastuzumab Treatment Holds Attributed to Symptomatic or Asymptomatic Cardiotoxicity
Trastuzumab was held for either asymptomatic declines in LVEF (Table 1) or possible cardiac symptoms. Thirty-one percent of evaluable patients temporarily held or permanently discontinued trastuzumab because of asymptomatic declines in LVEF (24%) and/or symptomatic cardiotoxicity (8%).
Cardiac Risk Factors and Incidence of CHF
Table 4 lists potential risk factors for CHF. LVEF, assessed either at baseline or after AC, was strongly associated with subsequent CHF (P < .0001), and age at entry was also predictive (P = .03). Hypertension was marginally significant (P = .07). In a multivariate analysis, age and post-AC LVEF remained statistically significant. There was no increase in CHF among women with left-sided lesions receiving radiotherapy (HR = 0.80; P = .59). Table 5 lists the3-year cumulative incidences of CHF among trastuzumab-treated patients as a function of post-AC LVEF and age. Results accorded with expectations for all patient groups except patients whose post-AC LVEF was marginally greater than LLN (ie, 54%). Among the 48 patients of age 50 years with LVEF in this range, nine experienced CHF, all within 7 months of initiating trastuzumab (cumulative incidence = 20%; 95% CI, 11% to 36%).
Changes in LVEF During Treatment
Patients accrued by June 30, 2003 have completed all MUGA scans through 18 months. In this group, the median baseline LVEF was 63%, and the median decrease after AC was 2% (P < .0001). The post-AC median decreases in the two arms were not statistically different (P = .25). In evaluable patients, there was an additional decrease in LVEF after paclitaxel with or without trastuzumab. Decrements were greater in arm 2 than arm 1. At 6, 9, and 18 months, the median declines in LVEF from baseline values were 5%, 6%, and 4% in arm 2 compared with 3%, 2%, and 3% in arm 1. At each point, the decline was statistically significant compared with baseline (P < .0001). The decline in arm 2 after the initiation of trastuzumab was greater than in arm 1 (at 6 and 9 months, P < .0001; at 18 months, P = .01).
CREC Criteria
We computed the cumulative incidence of MUGA results meeting CREC criteria for asymptomatic CD (decline in LVEF 10 percentage points to < 55%) at least once during the 52 weeks beginning with day 1 of cycle 5. In the evaluable cohort, the incidence was 17% (95% CI, 15% to 20%) in arm 1 and 34% (95% CI, 31% to 38%) in arm 2 (HR = 2.1; 95% CI, 1.7 to 2.6; P < .0001).
DISCUSSION
It is reasonable to compare the 4.1% incidence of CHF in arm 2 with the 2.2% incidence of class III or IV CHF in metastatic patients treated with paclitaxel and trastuzumab in the CREC analysis.3 However, comparison of incidence of CD that did not meet CHF criteria is difficult. Although 34% of B-31 arm 2 patients met CREC criteria for asymptomatic CD at least once during the 52 weeks after initiation of trastuzumab, the same was true for 17% of control patients. This rate of subclinical CD on the control arm was higher than the 13.2% rate of paclitaxel-trastuzumab–treated patients with metastatic cancer in the pivotal trial with any degree of CD.3 The requirement for scheduled serial MUGA scans and the prospective reporting requirements in B-31 complicate comparison of rates of overall CD.
To more clearly quantify rates of overall CD, it may be more appropriate to consider only asymptomatic declines in LVEF that were persistent and thus required discontinuation of trastuzumab. Of the 714 arm 2 patients for whom complete treatment data are available, 135 (19%) discontinued trastuzumab because of asymptomatic or symptomatic CD (n = 133, including 29 patients with CD meeting criteria for class III or IV CHF) or completed 1 year of treatment but subsequently presented with CHF (n = 2). This percentage (19%) may be a reasonable summary of arm 2 patients experiencing some degree of CD, including but not limited to patients who experienced class III or IV CHF.
Whether a 4.1% incidence of class III or IV CHF and an overall CD incidence of 19% are acceptable in the adjuvant setting depends on the magnitude of benefit from the therapy. Based on the results of the joint analysis of B-31 and NCCTG N9831, this benefit is considerable. At 4 years, the estimated absolute improvement in disease-free survival is 18% (85% minus 67%; 95% CI, 13% to 24%), and, for overall survival, the benefit is 4.8% (91.4% minus 86.6%; 95% CI, 0.6% to 9.0%).4 Another important consideration is reversibility of CD. The CHF associated with trastuzumab seemed to be responsive to cessation of therapy and management (Table 2), and LVEF generally recovered to nearly normal levels over time (Figs 3A and 3B). These findings suggest that CD associated with trastuzumab is qualitatively different than anthracycline cardiotoxicity.8 However, at present, patients with CHF associated with trastuzumab therapy have been followed for, at most, several years, and longer follow-up is required to understand the impact of this toxicity on survival and quality of life. Similarly, patients experiencing milder symptomatic or asymptomatic CD also need to be followed to fully determine the long-term implications.
To address risk-benefit trade-offs, it is essential to evaluate potential predictive factors for increased risk of toxicity. Post-AC LVEF and age were associated with trastuzumab-associated CHF. Age was a significant predictor of cardiac risk in the CREC analysis3 and is also a risk factor for doxorubicin-induced CD.9
Trastuzumab has been evaluated in the adjuvant setting using less cardiotoxic regimens than those in B-31. Recent data from NCCTG N9831 indicated delaying administration of trastuzumab until completion of paclitaxel may result in lower rates of severe CD.10 However, an unplanned interim efficacy analysis of the N9831 data suggested that this approach may be less effective than concurrent administration.10 Conversely, the Herceptin Adjuvant (HERA) trial, which compared trastuzumab to observation after completion of chemotherapy, resulted in little symptomatic CD and demonstrated improvement in disease-free survival comparable to the disease-free survival seen in B-31/N9831.11 Currently, Breast Cancer International Research Group Study 006 is evaluating trastuzumab administered concurrently with docetaxel and carboplatin. Toxicity data from this trial indicate minimal CD.12 If efficacy results are ultimately favorable, this study could support removal of anthracyclines from the chemotherapy regimen with which trastuzumab will be administered.
Given the marked reduction in recurrence and mortality afforded by the administration of trastuzumab with paclitaxel after AC, an assessment of risks and benefits indicates that the regimen is appropriate for most women with HER2-positive node-positive disease and adequate cardiac function. The regimen may not be appropriate for older patients whose post-AC LVEF is only marginally greater than LLN.
Authors' Disclosures of Potential Conflicts of Interest
Although all authors completed the disclosure declaration, the following authors or their immediate family members indicated a financial interest. No conflict exists for drugs or devices used in a study if they are not being evaluated as part of the investigation. For a detailed description of the disclosure categories, or for more information about ASCO’s conflict of interest policy, please refer to the Author Disclosure Declaration and the Disclosures of Potential Conflicts of Interest section in Information for Contributors.
Acknowledgment
We thank the women who participated in this study and the National Surgical Adjuvant Breast and Bowel Project Trial (NSABP) investigators who carried out the trial. We also thank Barbara C. Good, PhD, Director of Scientific Publications for the NSABP, for editorial assistance with this manuscript.
NOTES
Supported by Public Health service Grant Nos. U10-CA-12027 and U10-CA-69651 from the National Cancer Institute, National Institutes of Health, Department of Health and Human Services, Bethesda, MD, and by Genentech Inc, South San Francisco, CA.
Presented in part at the 26th Annual San Antonio Breast Cancer Symposium, San Antonio, TX, December 3-6, 2003.
Authors' disclosures of potential conflicts of interest are found at the end of this article.
REFERENCES
Slamon DJ, Clark GM, Wong SG, et al: Human breast cancer: Correlation of relapse and survival with amplification of the HER2/neu oncogene. Science 235:177-182, 1987
Slamon DJ, Leyland-Jones B, Shak S, et al: Use of chemotherapy plus a monoclonal antibody against HER2 for metastatic breast cancer that overexpresses HER2. N Engl J Med 344:783-792, 2001
Seidman A, Hudis Pierri MK, et al: Cardiac dysfunction in the trastuzumab clinical trials experience. J Clin Oncol 20:1215-1221, 2002
Romond EH, Perez EA, Bryant J, et al: Trastuzumab plus adjuvant chemotherapy for operable HER2-positive breast cancer: Combined analysis of NSABP B-31 and NCCTG N9831. N Engl J Med 353:15-26, 2005
Korn EL, Dorey FJ: Application of crude incidence curves. Stat Med 11:813-829, 1992
Peto R, Peto J: Asymptotically efficient rank invariant test procedures. J R Statist Soc A 135:185-207, 1972
Geyer CE Jr, Bryant J, Romond E, et al: Cardiac safety analysis of the first stage of NSABP B-31, a randomized trial comparing the safety and efficacy of Adriamycin and cyclophosphamide followed by Taxol plus Herceptin in patients with operable node-positive, HER2-overexpressing breast cancer (HER2+BC). 26th Annual San Antonio Breast Cancer Symposium, San Antonio, TX, December 3-6, 2003 (abstr 23)
Ewer MS, Lippman SM: Type II chemotherapy-related cardiac dysfunction: Time to recognize a new entity. J Clin Oncol 23:2900-2903, 2005
Swain S, Whaley F, Ewer M: Congestive heart failure in patients treated with doxorubicin: A retrospective analysis of three trials. Cancer 97:2869-2879, 2003
Perez EA, Suman VJ, Davidson N, et al: NCCTG N9831: May 2005 update. American Society of Clinical Oncology Scientific Symposia, Special Education Session, Orlando, FL, May 16, 2005
Piccart-Gebhart M: First results of the HERA trial. American Society of Clinical Oncology Scientific Symposia, Special Education Session, Orlando, FL, May 16, 2005
Weng W-K, Arteaga C, Slamon D: Antibody–based therapeutics: More than a one-trick pony (SSO2A). American Society of Clinical Oncology Scientific Symposia, Special Education Sessions, Orlando, FL, May 14-17, 2005(Elizabeth Tan-Chiu, Greg )
NSABP Biostatistical Center, University of Pittsburgh
Allegheny General Hospital
University of Pittsburgh Cancer Institute, Breast Cancer Institute, and Magee Women's Hospital, Pittsburgh, PA
Cancer Research Network, Inc, Plantation, FL
University of Kentucky Markey Cancer Center, Lexington, KY
The University of Texas, M.D. Anderson Cancer Center, Houston, TX
Novartis Pharmaceutical Corporation, East Hanover, NJ
Cancer Therapeutics Branch, National Cancer Institute, Bethesda, MD
Kaiser Permanente Vallejo, Vallejo, CA
Atlanta Cancer Care, Atlanta, GA
Aultman Hospital, Canton, OH.
ABSTRACT
PURPOSE: Trastuzumab is effective in treating human epidermal growth factor receptor 2 (HER2) –positive breast cancer, but it increases frequency of cardiac dysfunction (CD) when used with or after anthracyclines.
PATIENTS AND METHODS: National Surgical Adjuvant Breast and Bowel Project trial B-31 compared doxorubicin and cyclophosphamide (AC) followed by paclitaxel with AC followed by paclitaxel plus 52 weeks of trastuzumab beginning concurrently with paclitaxel in patients with node-positive, HER2-positive breast cancer. Initiation of trastuzumab required normal post-AC left ventricular ejection fraction (LVEF) on multiple-gated acquisition scan. If symptoms suggestive of congestive heart failure (CHF) developed, source documents were blindly reviewed by an independent panel of cardiologists to determine whether criteria were met for a cardiac event (CE), which was defined as New York Heart Association class III or IV CHF or possible/probable cardiac death. Frequencies of CEs were compared between arms.
RESULTS: Among patients with normal post-AC LVEF who began post-AC treatment, five of 814 control patients subsequently had confirmed CEs (four CHFs and one cardiac death) compared with 31 of 850 trastuzumab-treated patients (31 CHFs and no cardiac deaths). The difference in cumulative incidence at 3 years was 3.3% (4.1% for trastuzumab-treated patients minus 0.8% for control patients; 95% CI, 1.7% to 4.9%). Twenty-seven of the 31 patients in the trastuzumab arm have been followed for 6 months after diagnosis of a CE; 26 were asymptomatic at last assessment, and 18 remained on cardiac medication. CHFs were more frequent in older patients and patients with marginal post-AC LVEF. Fourteen percent of patients discontinued trastuzumab because of asymptomatic decreases in LVEF; 4% discontinued trastuzumab because of symptomatic cardiotoxicity.
CONCLUSION: Administering trastuzumab with paclitaxel after AC increases incidence of CHF and lesser CD. Potential cardiotoxicity should be carefully considered when discussing benefits and risks of this therapy.
INTRODUCTION
Trastuzumab (Herceptin; Genentech, South San Francisco, CA) is a humanized monoclonal antibody targeting the extracellular domain of the human epidermal growth factor receptor 2 (HER2) protein. Amplification or overexpression of HER2 is seen in approximately 25% of invasive breast cancers and is associated with unfavorable prognosis.1 Trastuzumab was approved for treatment of HER2-positive metastatic breast cancer in 1998 based on improved outcomes in the pivotal trial that evaluated trastuzumab in combination with chemotherapy. However, an increased incidence of cardiac dysfunction (CD) was observed, particularly when trastuzumab was combined with doxorubicin/cyclophosphamide (AC).2
As part of the process of filing for US Food and Drug Administration approval for use in metastatic breast cancer, an independent Cardiac Review and Evaluation Committee (CREC) was convened to review the cardiotoxicity data from all metastatic disease trials evaluating trastuzumab and established the following criteria for diagnosis of CD: (1) cardiomyopathy characterized by a global decrease in left ventricular ejection fraction (LVEF); (2) signs or symptoms of congestive heart failure (CHF); (3) decline in LVEF of at least 5% to less than 55% with signs or symptoms of CHF or decline in LVEF of at least 10% to less than 55% without signs or symptoms of CHF. Episodes of CD were characterized according to the New York Heart Association (NYHA) functional classification.3 AC alone was associated with an overall rate of CD of 8.1% and a class III or IV rate of 3.7%. Addition of trastuzumab to AC was associated with an increase in overall CD rate to 27.3% and an increase in the rate of class III or IV CD to 16.1%. Administration of trastuzumab with paclitaxel in patients with prior anthracycline exposure resulted in an overall CD rate of 13.2% but a class III/IV rate of only 2.2%. The CREC also reviewed patients after treatment for CD when possible. Of 34 patients who developed CD on AC plus trastuzumab, seven had persistent class III or IV symptoms on follow-up compared with 0 of 10 patients who developed CD on paclitaxel plus trastuzumab.3
The pivotal trial results justified evaluating trastuzumab with chemotherapy in the HER2-positive adjuvant patient population. This required a trial design that would minimize risk and evaluate the incidence, severity, and reversibility of CD. The low rate and apparent reversibility of class III or IV CHF that was observed when trastuzumab was combined with paclitaxel supported the comparison of AC followed by paclitaxel ± trastuzumab.
Therefore, National Surgical Adjuvant Breast and Bowel Project (NSABP) trial B-31 included two stages, with assessment of CD as the primary aim of the first stage. Stage 1 included the first 1,000 patients who received post-AC treatment and were evaluable for cardiac assessment. After favorable review of the stage 1 cardiac data by an independent data monitoring committee (DMC), accrual was extended to the second stage to enroll an additional 1,700 patients. A recent planned joint interim analysis of B-31 and North Central Cancer Treatment Group (NCCTG) N9831 demonstrated a 52% reduction in hazard of first events (recurrence, second primary, or death; hazard ratio [HR] = 0.48; 95% CI, 0.39 to 0.59; P = 2 x 10–12) and a 33% reduction in hazard of mortality (HR = 0.67; 95% CI, 0.48 to 0.93; P = .015), prompting the DMCs of both groups to recommend disclosure.4 We report here the detailed cardiac safety data from B-31 as of that analysis.
PATIENTS AND METHODS
Patient Eligibility and Entry Procedures
Patients must have had histologically node-positive HER2-positive primary breast cancer without evidence of metastatic disease and must have completed primary surgery and axillary dissection. LVEF on multiple-gated acquisition (MUGA) scan had to be the lower limit of normal (LLN). Patients with active cardiac disease, including angina pectoris or cardiac arrhythmia requiring medication, severe conduction abnormalities, clinically significant valvular disease, cardiomegaly, ventricular hypertrophy, or poorly controlled hypertension, were ineligible. Patients with prior myocardial infarction, CHF, or other cardiomyopathy were also excluded. Protocol approval by the local human investigations committee or IRB was required, with an assurance filed with the Department of Health and Human Services. Written informed consent was required.
Protocol Treatment Regimens
Regimens consisted of either doxorubicin (60 mg/m2) and cyclophosphamide (600 mg/m2) every 21 days for four cycles followed by paclitaxel (175 mg/m2) every 3 weeks for four cycles (arm 1) or the same chemotherapy plus weekly trastuzumab starting with the first dose of paclitaxel at a loading dose of 4 mg/kg followed by 2 mg/kg for 51 weeks (arm 2). (Subsequent to May 16, 2003, paclitaxel could be administered weekly at 80 mg/m2 for 12 weeks.) Patients with lumpectomies received breast irradiation after chemotherapy. Postlumpectomy regional and postmastectomy locoregional irradiation were optional. Trastuzumab was continued during radiation. Radiation of internal mammary nodes was prohibited. Women with hormone receptor–positive tumors were to receive tamoxifen for 5 years.
Cardiac Monitoring
Cardiac history forms were to be submitted at entry, every 6 months for the first 5 years, and yearly thereafter. MUGA scans were required in both treatment arms before entry, after AC, and at 6, 9, and 18 months. Additional scans could be performed at investigator discretion.
Initiation of Trastuzumab and Treatment Discontinuation
Trastuzumab was initiated in arm 2 patients if they remained without cardiac symptoms after AC and their LVEF remained LLN and had not decreased by more than 15 percentage points from the pre-entry level. Criteria were established for holding trastuzumab in asymptomatic patients if 6- and 9-month LVEF values did not remain within predefined limits, as listed in Table 1. If a participant's LVEF did not meet criteria for continuing trastuzumab, the drug was held for 4 weeks. If a repeat LVEF did not meet criteria for continuation, further administration was prohibited.
Monitoring and Review of Symptomatic CD
We will use the term CHF to denote NYHA class III or IV CD. Patients with lesser dysfunction are not referred to as having experienced CHF, although that term is sometimes used to refer to lesser degrees of CD in routine clinical practice.
All symptoms suggestive of CHF were reported within 14 days. Source documents were blinded to treatment arm and reviewed by an independent Cardiac Review Panel consisting of three cardiologists who determined by majority vote whether each patient met prospectively defined criteria for CHF. Criteria mirrored the CREC criteria as follows: NYHA class III or IV symptoms with either a decrease from baseline in LVEF of more than 10 percentage points to less than 55% or a decrease of more than 5 percentage points to less than the LLN. Class III is characterized by symptoms of dyspnea with activities such as climbing a flight of stairs, whereas class IV symptoms are present at rest.
Patients for whom potential CD was reported continued to be monitored by MUGA scans and cardiac follow-up forms that summarized ongoing symptoms and cardiac medications.
Formal Interim Analysis of Cardiotoxicity
The primary cardiac safety end point was the difference in incidence of cardiac events (CE) between arms, defined as confirmed class III or IV CHF or possible/probable cardiac death. During study planning, it was judged that at least a 25% reduction in hazard of mortality was not unlikely with this treatment. This translates to an 8% absolute survival benefit at 10 years (70% v 62%). Given this level of expected benefit, there was agreement that a difference in incidence of CE of up to 4% would be acceptable to allow completion of accrual in stage 2. Cardiac interim analyses were scheduled after 200, 600, and 1,000 v patients had been followed for at least 6 months after the first day of treatment with paclitaxel ± trastuzumab. If at any interim analysis the difference in CE occurrence was statistically greater than 4% (one-sided P < .05), accrual was to be suspended. Otherwise, accrual would continue to a target of 2,700 patients to assess efficacy. This interim analysis plan had a 12% probability of terminating accrual if the true difference in event rate was 4% and an 85% probability of terminating accrual if the difference was 7%.
Statistical Methods
This report is based on all patients accrued as of February 15, 2005 and all cardiac follow-up data received by April 21, 2005. Primary comparisons are based on the cohort of patients who completed AC and began post-AC treatment with paclitaxel (arm 1) or trastuzumab plus paclitaxel (arm 2). To ensure comparability, all patients in both arms of this evaluable cohort were required to meet the protocol conditions for initiating trastuzumab that were required of arm 2 patients (ie, patients were without cardiac symptoms after AC and had a post-AC LVEF LLN with a 15 percentage points decline from baseline). Patients in this cohort were considered to be at risk for CE from day 1 of cycle 5 until recurrence, diagnosis of second primary cancer, or last follow-up. Cumulative incidence was estimated nonparametrically.5 The difference in cumulative incidence of CE in evaluable arm 1 and arm 2 patients 3 years after day 1 of cycle 5 was compared with the hypothesized difference of 4% using a one-sided 0.05-level test. The ratio of crude hazards for CE was estimated by fitting a Cox model. Secondary intent-to-treat analyses were performed using the cohort of all patients with cardiac follow-up.
Association of risk factors with CE was tested using univariate log-rank analysis. A permutation method was used to compute P values.6 In exploratory analyses, Cox models were fit to assess the independence of predictors. Changes in LVEF from baseline were tested using the signed rank procedure; comparisons of LVEF between arms were tested using the rank sum test applied to changes from baseline.
RESULTS
Accrual and Patient Characteristics
The trial opened on February 21, 2000. Cardiotoxicity interim analyses were reviewed by the DMC on October 1, 2001, October 11, 2002, and October 10, 2003. In all three analyses, the difference in CE rates was less than 4%, so patient accrual continued.7 By February 15, 2005, 2,043 patients had been enrolled. Figure 1 summarizes their treatment, evaluation, and follow-up. Twenty-two patients refused protocol therapy; AC treatment information is pending for 49 patients; 148 patients did not meet the post-AC cardiac criteria for initiation of trastuzuma and were excluded from the primary cardiac safety analysis, regardless of assigned treatment arm; 36 patients had pending post-AC MUGA scans; and 124 patients did not begin post-AC therapy or had no post-AC follow-up as of this report. The remaining 1,664 patients (81%) are included in the evaluable cohort. The median follow-up of patients in the evaluable cohort is 27 months from the first day of cycle 5. Characteristics of evaluable patients are balanced by treatment arm. Fifty-one percent of patients were 49 years old at entry, 33% were 50 to 59 years old, and 16% were 60 years old. Eighty-four percent of patients were white, 8% were black, and 8% were of other or unknown race/ethnicity. Fifty-seven percent of patients had one to three positive nodes, 29% had four to nine positive nodes, and 14% had 10 positive nodes. Fifty-nine percent of patients had a pathologic tumor size of more than 2.0 cm. Fifty-two percent of patients had estrogen receptor–positive tumors.
Cumulative Incidence of CEs
Among 814 evaluable control patients, five met criteria for a CE (four CHFs and one probable cardiac death). In the arm 2 evaluable cohort, 31 of 850 patients experienced a CE (31 CHFs and no cardiac deaths). Cumulative incidence of CEs in the control arm 3 years after day 1 of cycle 5 was 0.8% (95% CI, 0.3% to 1.9%). Among trastuzumab-treated evaluable patients, the cumulative incidence of CEs 3 years after day 1 of cycle 5 was 4.1% (95% CI, 2.9% to 5.8%; Fig 2). The difference between the arms was 3.3% (95% CI, 1.7% to 4.9%), which was in agreement with the 4% anticipated increase (P > .5). The relative risk of a CE was 5.9 in trastuzumab patients versus control patients (95% CI, 2.3 to 15.3; P < .0001). Three of the 31 cases of CHF in arm 2 occurred more than 1 year after initiation of trastuzumab.
Five additional CEs occurred in patients not included in the evaluable cohort (three patients in arm 1 and two in arm 2; Fig 1). In arm 2, a patient died of CHF after her third cycle of AC. A second patient with multiple cardiac risk factors died of cardiopulmonary collapse after her first dose of AC. Neither patient received paclitaxel or trastuzumab. In arm 1, three patients with post-AC LVEF less than LLN received paclitaxel and subsequently experienced CHF. (These patients were not protocol violations because normal post-AC LVEF was not required to initiate paclitaxel, although it was required in arm 2 patients to receive trastuzumab.) The intent-to-treat 3-year cumulative incidence of CEs in all arm 1 patients was 1.1% (95% CI, 0.6% to 2.3%) compared with 3.9% in the entire arm 2 cohort (95% CI, 2.8% to 5.5%).
Recovery of Cardiac Function After CHF
Among the 31 evaluable patients meeting criteria for CHF who received trastuzumab, none have died for reasons other than metastatic breast cancer. Twenty-seven of the 31 patients have been followed for 6 months after diagnosis of CHF, of whom 26 have been without symptoms of CD for at least 6 months. However, 18 patients continued to receive cardiac medications (Table 2). Twenty-four of the 27 patients have reported LVEF assessments 6 months after the diagnosis of their CHF, and of these 24 patients, 17 had decreased LVEF relative to baseline at their last available MUGA. The most recent LVEF values were between 40% and 69% in all except one patient whose LVEF was 25% (Fig 3B). Of the five confirmed patients with CHF in arm 1, three have died (probable cardiac death, metastatic breast cancer, emphysema). One of the two surviving patients has been followed for 6 months after her diagnosis of CHF and is currently symptomatic and receiving medications (Table 2).
Recovery of Cardiac Function in Patients With Symptoms of Possible CD That Did Not Meet Criteria for CE
In addition to those patients with confirmed CHF, 43 evaluable arm 2 patients reported symptoms of CD that did not meet protocol criteria for a CE. Thirty-nine patients have been followed for 6 months since onset of symptoms, of whom one patient reported symptoms of CD at last follow-up; eight patients continued to receive cardiac medication (Table 2). MUGA results have been reported 6 months after onset of symptoms for 30 of the 39 patients. Current LVEF results range from 45% to 86%, except for one patient whose LVEF was 40% (Fig 3D). Twenty-two of these patients had decreased LVEF relative to baseline at last evaluation, suggesting some decrement in function.
In addition to the five evaluable arm 1 patients with confirmed CEs, eight evaluable arm 1 patients reported symptoms of CD before recurrence that did not meet protocol criteria for a CE. Six patients have been followed for 6 months since onset of symptoms; all patients were asymptomatic and receiving no cardiac medications (Table 2).
Discontinuation of Trastuzumab Therapy in Patients in Arm 2
Trastuzumab treatment information was submitted quarterly. At least one treatment form has been submitted on 825 of 850 evaluable arm 2 patients, and 714 have completed trastuzumab, and all required documentation has been submitted. Trastuzumba was permanently discontinued before completion of 1 year of therapy or recurrence in 197 patients (28%) (Table 3). Trastuzumab was discontinued for cardiac-related reasons (asymptomatic decline in LVEF, symptoms of CD, or other cardiac symptoms) in 133 patients (19%), including 17 (2%) of 825 patients during the first quarter of therapy, 57 (8%) of 757 patients during the second quarter, 44 (7%) of 643 patients during the third quarter, and 15 (3%) of 538 patients during the final quarter.
Of 102 evaluable patients who discontinued trastuzumab because of asymptomatic declines in LVEF, 21 subsequently reported symptoms of possible CD and have been included in the previous cohorts. Of the remainder, LVEF has been reported in 52 patients 6 months after discontinuation. Their nadir and most recent LVEF results are shown in Figures 3E and 3F, where they are contrasted with the nadir and most recent results for trastuzumab-treated patients with confirmed CEs (Figs 3A and 3B) and patients who had symptoms of CD not meeting criteria for a CE (Figs 3C and 3D). Of these 52 patients, the most current LVEFs are less than 45% in three patients (6%) and less than 50% in 13 patients (25%).
Trastuzumab Treatment Holds Attributed to Symptomatic or Asymptomatic Cardiotoxicity
Trastuzumab was held for either asymptomatic declines in LVEF (Table 1) or possible cardiac symptoms. Thirty-one percent of evaluable patients temporarily held or permanently discontinued trastuzumab because of asymptomatic declines in LVEF (24%) and/or symptomatic cardiotoxicity (8%).
Cardiac Risk Factors and Incidence of CHF
Table 4 lists potential risk factors for CHF. LVEF, assessed either at baseline or after AC, was strongly associated with subsequent CHF (P < .0001), and age at entry was also predictive (P = .03). Hypertension was marginally significant (P = .07). In a multivariate analysis, age and post-AC LVEF remained statistically significant. There was no increase in CHF among women with left-sided lesions receiving radiotherapy (HR = 0.80; P = .59). Table 5 lists the3-year cumulative incidences of CHF among trastuzumab-treated patients as a function of post-AC LVEF and age. Results accorded with expectations for all patient groups except patients whose post-AC LVEF was marginally greater than LLN (ie, 54%). Among the 48 patients of age 50 years with LVEF in this range, nine experienced CHF, all within 7 months of initiating trastuzumab (cumulative incidence = 20%; 95% CI, 11% to 36%).
Changes in LVEF During Treatment
Patients accrued by June 30, 2003 have completed all MUGA scans through 18 months. In this group, the median baseline LVEF was 63%, and the median decrease after AC was 2% (P < .0001). The post-AC median decreases in the two arms were not statistically different (P = .25). In evaluable patients, there was an additional decrease in LVEF after paclitaxel with or without trastuzumab. Decrements were greater in arm 2 than arm 1. At 6, 9, and 18 months, the median declines in LVEF from baseline values were 5%, 6%, and 4% in arm 2 compared with 3%, 2%, and 3% in arm 1. At each point, the decline was statistically significant compared with baseline (P < .0001). The decline in arm 2 after the initiation of trastuzumab was greater than in arm 1 (at 6 and 9 months, P < .0001; at 18 months, P = .01).
CREC Criteria
We computed the cumulative incidence of MUGA results meeting CREC criteria for asymptomatic CD (decline in LVEF 10 percentage points to < 55%) at least once during the 52 weeks beginning with day 1 of cycle 5. In the evaluable cohort, the incidence was 17% (95% CI, 15% to 20%) in arm 1 and 34% (95% CI, 31% to 38%) in arm 2 (HR = 2.1; 95% CI, 1.7 to 2.6; P < .0001).
DISCUSSION
It is reasonable to compare the 4.1% incidence of CHF in arm 2 with the 2.2% incidence of class III or IV CHF in metastatic patients treated with paclitaxel and trastuzumab in the CREC analysis.3 However, comparison of incidence of CD that did not meet CHF criteria is difficult. Although 34% of B-31 arm 2 patients met CREC criteria for asymptomatic CD at least once during the 52 weeks after initiation of trastuzumab, the same was true for 17% of control patients. This rate of subclinical CD on the control arm was higher than the 13.2% rate of paclitaxel-trastuzumab–treated patients with metastatic cancer in the pivotal trial with any degree of CD.3 The requirement for scheduled serial MUGA scans and the prospective reporting requirements in B-31 complicate comparison of rates of overall CD.
To more clearly quantify rates of overall CD, it may be more appropriate to consider only asymptomatic declines in LVEF that were persistent and thus required discontinuation of trastuzumab. Of the 714 arm 2 patients for whom complete treatment data are available, 135 (19%) discontinued trastuzumab because of asymptomatic or symptomatic CD (n = 133, including 29 patients with CD meeting criteria for class III or IV CHF) or completed 1 year of treatment but subsequently presented with CHF (n = 2). This percentage (19%) may be a reasonable summary of arm 2 patients experiencing some degree of CD, including but not limited to patients who experienced class III or IV CHF.
Whether a 4.1% incidence of class III or IV CHF and an overall CD incidence of 19% are acceptable in the adjuvant setting depends on the magnitude of benefit from the therapy. Based on the results of the joint analysis of B-31 and NCCTG N9831, this benefit is considerable. At 4 years, the estimated absolute improvement in disease-free survival is 18% (85% minus 67%; 95% CI, 13% to 24%), and, for overall survival, the benefit is 4.8% (91.4% minus 86.6%; 95% CI, 0.6% to 9.0%).4 Another important consideration is reversibility of CD. The CHF associated with trastuzumab seemed to be responsive to cessation of therapy and management (Table 2), and LVEF generally recovered to nearly normal levels over time (Figs 3A and 3B). These findings suggest that CD associated with trastuzumab is qualitatively different than anthracycline cardiotoxicity.8 However, at present, patients with CHF associated with trastuzumab therapy have been followed for, at most, several years, and longer follow-up is required to understand the impact of this toxicity on survival and quality of life. Similarly, patients experiencing milder symptomatic or asymptomatic CD also need to be followed to fully determine the long-term implications.
To address risk-benefit trade-offs, it is essential to evaluate potential predictive factors for increased risk of toxicity. Post-AC LVEF and age were associated with trastuzumab-associated CHF. Age was a significant predictor of cardiac risk in the CREC analysis3 and is also a risk factor for doxorubicin-induced CD.9
Trastuzumab has been evaluated in the adjuvant setting using less cardiotoxic regimens than those in B-31. Recent data from NCCTG N9831 indicated delaying administration of trastuzumab until completion of paclitaxel may result in lower rates of severe CD.10 However, an unplanned interim efficacy analysis of the N9831 data suggested that this approach may be less effective than concurrent administration.10 Conversely, the Herceptin Adjuvant (HERA) trial, which compared trastuzumab to observation after completion of chemotherapy, resulted in little symptomatic CD and demonstrated improvement in disease-free survival comparable to the disease-free survival seen in B-31/N9831.11 Currently, Breast Cancer International Research Group Study 006 is evaluating trastuzumab administered concurrently with docetaxel and carboplatin. Toxicity data from this trial indicate minimal CD.12 If efficacy results are ultimately favorable, this study could support removal of anthracyclines from the chemotherapy regimen with which trastuzumab will be administered.
Given the marked reduction in recurrence and mortality afforded by the administration of trastuzumab with paclitaxel after AC, an assessment of risks and benefits indicates that the regimen is appropriate for most women with HER2-positive node-positive disease and adequate cardiac function. The regimen may not be appropriate for older patients whose post-AC LVEF is only marginally greater than LLN.
Authors' Disclosures of Potential Conflicts of Interest
Although all authors completed the disclosure declaration, the following authors or their immediate family members indicated a financial interest. No conflict exists for drugs or devices used in a study if they are not being evaluated as part of the investigation. For a detailed description of the disclosure categories, or for more information about ASCO’s conflict of interest policy, please refer to the Author Disclosure Declaration and the Disclosures of Potential Conflicts of Interest section in Information for Contributors.
Acknowledgment
We thank the women who participated in this study and the National Surgical Adjuvant Breast and Bowel Project Trial (NSABP) investigators who carried out the trial. We also thank Barbara C. Good, PhD, Director of Scientific Publications for the NSABP, for editorial assistance with this manuscript.
NOTES
Supported by Public Health service Grant Nos. U10-CA-12027 and U10-CA-69651 from the National Cancer Institute, National Institutes of Health, Department of Health and Human Services, Bethesda, MD, and by Genentech Inc, South San Francisco, CA.
Presented in part at the 26th Annual San Antonio Breast Cancer Symposium, San Antonio, TX, December 3-6, 2003.
Authors' disclosures of potential conflicts of interest are found at the end of this article.
REFERENCES
Slamon DJ, Clark GM, Wong SG, et al: Human breast cancer: Correlation of relapse and survival with amplification of the HER2/neu oncogene. Science 235:177-182, 1987
Slamon DJ, Leyland-Jones B, Shak S, et al: Use of chemotherapy plus a monoclonal antibody against HER2 for metastatic breast cancer that overexpresses HER2. N Engl J Med 344:783-792, 2001
Seidman A, Hudis Pierri MK, et al: Cardiac dysfunction in the trastuzumab clinical trials experience. J Clin Oncol 20:1215-1221, 2002
Romond EH, Perez EA, Bryant J, et al: Trastuzumab plus adjuvant chemotherapy for operable HER2-positive breast cancer: Combined analysis of NSABP B-31 and NCCTG N9831. N Engl J Med 353:15-26, 2005
Korn EL, Dorey FJ: Application of crude incidence curves. Stat Med 11:813-829, 1992
Peto R, Peto J: Asymptotically efficient rank invariant test procedures. J R Statist Soc A 135:185-207, 1972
Geyer CE Jr, Bryant J, Romond E, et al: Cardiac safety analysis of the first stage of NSABP B-31, a randomized trial comparing the safety and efficacy of Adriamycin and cyclophosphamide followed by Taxol plus Herceptin in patients with operable node-positive, HER2-overexpressing breast cancer (HER2+BC). 26th Annual San Antonio Breast Cancer Symposium, San Antonio, TX, December 3-6, 2003 (abstr 23)
Ewer MS, Lippman SM: Type II chemotherapy-related cardiac dysfunction: Time to recognize a new entity. J Clin Oncol 23:2900-2903, 2005
Swain S, Whaley F, Ewer M: Congestive heart failure in patients treated with doxorubicin: A retrospective analysis of three trials. Cancer 97:2869-2879, 2003
Perez EA, Suman VJ, Davidson N, et al: NCCTG N9831: May 2005 update. American Society of Clinical Oncology Scientific Symposia, Special Education Session, Orlando, FL, May 16, 2005
Piccart-Gebhart M: First results of the HERA trial. American Society of Clinical Oncology Scientific Symposia, Special Education Session, Orlando, FL, May 16, 2005
Weng W-K, Arteaga C, Slamon D: Antibody–based therapeutics: More than a one-trick pony (SSO2A). American Society of Clinical Oncology Scientific Symposia, Special Education Sessions, Orlando, FL, May 14-17, 2005(Elizabeth Tan-Chiu, Greg )