- © 2005 by American Society of Clinical Oncology
Reversibility of Trastuzumab-Related cardiotoxicity: New Insights Based on Clinical Course and Response to Medical Treatment
- Michael S. Ewer,
- Mary T. Vooletich,
- Jean-Bernard Durand,
- Myrshia L. Woods,
- Joseph R. Davis,
- Vicente Valero and
- Daniel J. Lenihan
- From the Departments of Cardiology and Breast Medical Oncology, The University of Texas M.D. Anderson Cancer Center, Houston, TX
- Address reprint requests to Michael S. Ewer, MD, and Daniel J. Lenihan, MD, Department of Cardiology, Unit 43, The University of Texas M.D. Anderson Cancer Center, 1515 Holcombe Blvd, Houston, TX 77030; e-mail: mewer{at}mdanderson.org, dlenihan{at}mdanderson.org
Abstract
Purpose Trastuzumab is an important biologic agent with significant activity in breast cancers that overexpress the HER2/neu marker. However, trastuzumab is associated with cardiotoxicity that has not yet been fully explored. We present our experience with patients who developed trastuzumab-related cardiotoxicity.
Patients and Methods Over a 4-year period, 38 patients with HER2/neu–positive breast cancer were referred for suspected trastuzumab-related cardiotoxicity. All patients had previously received anthracycline-based chemotherapy.
Results After doxorubicin but before trastuzumab, the mean (± standard deviation) left ventricular ejection fraction (LVEF) was 0.61 ± 0.13, and the LVEF decreased to 0.43 ± 0.16 after trastuzumab (P < .0001). After withdrawal of trastuzumab, the LVEF increased to 0.56 ± 0.11. Mean time to recovery of LVEF was 1.5 months and was temporally associated with medical treatment in 32 (84%) of the 38 patients but occurred without treatment in six patients (16%). Increases in LVEF were noted in 37 of the 38 patients. Twenty-five of these patients were re-treated with trastuzumab; three patients had recurrent left ventricular dysfunction, but 22 patients (88%) did not. All re-treatment patients continued on their therapeutic regimen for heart failure when rechallenged with trastuzumab. Nine patients underwent endomyocardial biopsy. Ultrastructural changes were not seen.
Conclusion Patients who develop cardiotoxicity while receiving trastuzumab therapy generally improve on removal of the agent. The mechanism of trastuzumab-related cardiac dysfunction is different from that of anthracycline cardiotoxicity, in part, demonstrated by the absence of anthracycline-like ultrastructural changes. Reintroducing trastuzumab may be appropriate for some individuals who previously have experienced trastuzumab-related cardiac dysfunction.
INTRODUCTION
Breast cancer affects more than 200,000 women per year and results in 40,000 deaths annually, making it the fourth leading cause of death in women in the United States.1,2 For more than 30 years, anthracyclines have been considered mainstays of treatment for early and advanced breast cancer.3 Regimens that include anthracyclines have demonstrated a clear, albeit modest, therapeutic advantage over regimens that do not include anthracyclines.4,5 It has been recognized that a subset of approximately 20% to 30% of patients with breast cancer have tumors that have HER2/neu gene amplification.6 Women with malignancy in which the HER2/neu gene is overexpressed generally have a less favorable response to traditional chemotherapy, including anthracyclines, primarily because of an increased growth rate of malignant cells.7,8 In addition, they have shorter periods of overall and disease-free survival and experience a higher incidence of brain metastasis. HER2/neu-positive tumors respond favorably to trastuzumab, a monoclonal antibody that selectively targets the extracellular domain of human epidermal growth factor receptor 2 (HER2).9 As a result, in this subset of patients, trastuzumab is becoming established as the standard of therapy for patients with metastatic breast cancer and is being evaluated for efficacy in earlier stages of disease.10 Both anthracyclines and trastuzumab are associated with considerable cardiotoxicity.11 Anthracycline-associated cardiotoxicity is related to the cumulative dose of the agent and the method of administration, and a threshold corresponding to an incidence of approximately 5% of congestive heart failure (CHF) has been targeted as a cutoff point for standard administration.12 Although this 5% threshold was originally thought to correspond with a cumulative dose of approximately 550 mg/m2, more recent studies have suggested that 400 to 450 mg/m2 corresponds more closely.13
Substantial variability exists in the incidence of cardiotoxicity, which is related not only to the cumulative dose and duration of infusion but also to the specific anthracycline formulation.11,14 In the case of trastuzumab, cardiotoxicity is less predictable but has been reported to occur in up to 7% of patients when trastuzumab is used as a single agent.15,16 When trastuzumab is combined with an anthracycline, cardiotoxicity is notably increased and has been reported to occur, at least in a mild form, in up to 27% of patients.15 More severe CHF, such as New York Heart Association class III or IV, was initially reported in 16% of patients in the pivotal trials of patients treated with this combination.15 Furthermore, other agents frequently used to treat breast cancer, such as paclitaxel, are reported to result in a higher than expected incidence of cardiotoxicity in patients concomitantly or sequentially treated with trastuzumab. Therefore, a general consensus has emerged that the concurrent use of trastuzumab with an anthracycline should be avoided. Because of the considerable tumor effect of trastuzumab in HER2/neu-positive breast cancer and the potential importance of combining trastuzumab with an anthracycline for optimal oncologic efficacy, at least four major trials are presently underway to help define the risks and benefits of such combinations in the adjuvant treatment of breast cancer.16 However, these studies will leave unanswered several important questions regarding the clinical course of patients treated with both an anthracycline and trastuzumab. To our knowledge, none of the major trials have included myocardial biopsy correlations, and none have addressed the outcome of patients who develop cardiac dysfunction or the management of such patients.
We now report our experience with 38 patients who received an anthracycline before or during treatment with trastuzumab and who subsequently developed left ventricular (LV) dysfunction and/or clinical CHF. We describe the clinical course of these patients and our approach to managing their LV dysfunction. In addition, we provide information regarding those patients who underwent cardiac biopsy while experiencing LV dysfunction and also describe our clinical observations with a subgroup of patients who, after recovery of LV function and treatment with standard CHF therapy, were rechallenged with trastuzumab. We hope that this report will offer an evidence-based approach to the management and prognosis of trastuzumab-related cardiotoxicity.
PATIENTS AND METHODS
An institutional review board–approved retrospective review was performed of patients who were referred to the Department of Cardiology at The University of Texas M.D. Anderson Cancer Center over a 4-year period (from January 2000 to January 2004) for possible trastuzumab-related cardiotoxicity. These included both patients who were apparently asymptomatic and patients who had developed symptomatic CHF. All patients underwent an initial cardiac evaluation that included a baseline LV ejection fraction (LVEF) measurement before trastuzumab. Thirty-seven patients had received treatment with anthracyclines before the LVEF measurement. One patient had received both doxorubicin and trastuzumab concurrently. These patients were referred for evaluation of possible trastuzumab-related cardiac toxicity; they constituted a subset of a larger group of breast cancer patients who were HER2/neu positive and who had received trastuzumab at our institution. In the 38 patients identified in this report, it is not possible to estimate the overall incidence of trastuzumab-related LV dysfunction. Most of the patients reviewed here received their initial treatment at our center; however, a small number may have been referred to other centers for cardiologic evaluation. Other patients have been referred specifically to our center, having undergone their initial treatment at outside facilities. However, it is known that approximately 125 patients are treated with trastuzumab per year at The M.D. Anderson Cancer Center, and the large majority of these patients do not develop detectable cardiotoxicity. On the basis of these assumptions and limitations, we believe that approximately 4% of patients who receive doxorubicin followed by trastuzumab experience clinically relevant LV dysfunction, and these patients form the basis of this review. This suggests that the incidence of trastuzumab-associated LV dysfunction is lower than initially reported. Large-scale randomized trials that are underway presently will provide further insight into this question and perhaps provide a better prospective on both the incidence of cardiotoxicity and the incremental survival benefit of the sequential use of doxorubicin followed by trastuzumab.
In our patients, the LVEF was routinely measured by a multiple-gated acquisition scan using standard established technologies.17 In some cases, echocardiography was used, and the LVEF was determined by the method of disks.18 Thirty-five (92%) of the 38 patients had a baseline LVEF within the normal range (0.50 to 0.70). If an LVEF was reported as a range of ± 5%, the mean value was used in all calculations. Sixteen patients (42%) were evaluated for coronary artery disease (CAD) using stress testing, and eight of these patients underwent cardiac catheterization. Four patients had significant CAD. Nine (24%) of the 38 patients underwent endomyocardial biopsy,19 which was performed either to evaluate for anthracycline-associated cardiotoxicity or because a specific protocol required evaluation of cardiac biopsy specimens. Thirty-eight patients (37 women and one man) were evaluated. The mean age (± standard deviation) was 52 ± 15 years (range, 29 to 73 years). The identified comorbidities before the initiation of trastuzumab therapy are listed in Table 1. Four patients (11%) had established CAD, and 10 (26%) had hypertension, requiring ongoing therapy. Other CAD risk factors were also present; seven of the patients (18%) continued to smoke despite the cardiotoxic nature of their treatment. All patients are offered smoking cessation counseling at our institution. Additionally, eight patients (21%) had a significant family history of premature CAD. Only one patient had hyperlipidemia. All patients received prior anthracyclines, and 35 (92%) of the 38 patients were concomitantly treated with cyclophosphamide. Twenty-seven patients (71%) had a history of prior exposure to a taxane. The majority of patients (25 of 38 patients, 66%), because of a history of previous response to trastuzumab-based therapy, were ultimately rechallenged with trastuzumab once a period of stability or improvement in LV dysfunction was established. All of these patients received standard CHF medication (including an angiotensin-converting enzyme inhibitor and beta-blocker therapy) before and during the reintroduction of trastuzumab therapy. In addition, all patients were extensively counseled on the potential risks and benefits of re-treatment with trastuzumab.
Data for individual patients were compared using a paired, two-tailed t test, whereas data summarizing a group of patients were compared using repeated measures of analysis of variance, with a P < .05 considered significant in all comparisons.
RESULTS
The mean initial (after anthracycline and before trastuzumab) LVEF for the entire group was 0.61 ± 0.13, and only three patients had a baseline LVEF of less than 0.50. After 4.5 months (median) of trastuzumab therapy, the mean LVEF decreased to 0.43 ± 0.16 (P < .0001 v baseline LVEF by paired t test). Once LV dysfunction or CHF was identified, trastuzumab therapy was discontinued in 37 of the 38 patients. The patients were either treated for cardiac dysfunction (n = 31) or observed (n = 6). The one remaining patient had pre-existing LV dysfunction and CHF; for compassionate reasons and with informed consent, she was allowed to continue trastuzumab therapy.
During a period of approximately 1.5 months, the LVEF returned toward baseline, with the mean LVEF increasing to 0.55 ± 0.11 (P < .001 v post-trastuzumab LVEF by paired t test; Fig 1). The LVEF at 1 to 3 months improved in all of the patients except the one patient who continued on trastuzumab; she had no clinical progression of CHF and demonstrated a slight improvement in LVEF once treatment with standard CHF medications commenced (Fig 2). Although all of our 38 patients had decreased LVEF with trastuzumab, 20 of the patients (52%) also experienced symptomatic CHF. Fifteen of these patients (75%), along with 16 patients who did not experience CHF, were treated with standard CHF medications, including angiotensin-converting enzyme inhibitors and beta-blockers, whereas five patients with and one patient without symptoms were observed without specific treatment for their CHF (Fig 3). The decision to treat was based on the discretion of the treating physician. All of the 20 patients with CHF had some improvement in LVEF and significant symptomatic improvement when trastuzumab was discontinued; however, two of the five patients who were not treated with CHF medications had persistent LV dysfunction when reassessed approximately 6 months later.
Trastuzumab Re-Treatment
Once a period of stability in CHF symptoms and LVEF was achieved, 25 (66%) of the 38 patients were reintroduced to trastuzumab because they had a considerable initial response to trastuzumab. All 25 patients were informed of the potential risks and benefits and gave written informed consent to proceed. All of the patients received the maximum-tolerated doses of angiotensin-converting enzyme inhibitors and beta-blockers while being rechallenged with trastuzumab. The median duration of reintroduced trastuzumab therapy was 8.4 months. Twenty-two (88%) of the 25 patients had stable LVEF measurements once treatment with trastuzumab reinitiated without recurrences of CHF during follow-up. LV dysfunction and/or CHF reoccurred in the remaining three patients (12%), and trastuzumab therapy was permanently discontinued (Fig 3). None of the patients died as a result of cardiac disease, and only one patient was hospitalized because of CHF with volume overload.
Cardiac Biopsy
Nine patients underwent right ventricular endomyocardial biopsy per protocol or as part of their evaluation for new onset of severe LV dysfunction. None of these patients showed any evidence of typical anthracycline-related changes on cardiac biopsy material.19 Specifically, no vacuolization or myocyte dropout was observed (Fig 4).
DISCUSSION
The use of trastuzumab for patients with breast cancer that overexpresses the HER2/neu marker has significantly increased response rates and improved survival and is generally considered a major breakthrough in the treatment of breast cancer. Furthermore, a recent trial reported that, even at early stages of disease, trastuzumab-containing regimens resulted in complete early pathologic responses in a significant number of patients (> 65%) whose tumors were to be surgically resected.20 The principal findings of our observational study indicate that patients who experience cardiotoxicity while receiving trastuzumab therapy generally recover their cardiac function when trastuzumab is discontinued. This improvement usually occurs over a period of months (mean time, 1.5 months) after withdrawal of the agent. In contrast to anthracycline-related cardiotoxicity, which results in observable ultrastructural changes on the myocardial biopsy, trastuzumab-related cardiotoxicity did not create detectable ultrastructural abnormalities. Additionally, patients who experienced a benefit with trastuzumab and demonstrated a resolution or improvement of cardiotoxicity could, with acceptable risk, be re-treated with trastuzumab while receiving protective CHF medications and undergoing close cardiac function monitoring. Our data suggest that cardiotoxicity resulting from trastuzumab is largely reversible, unlike the cardiotoxicity associated with anthracyclines, and that trastuzumab need not be withheld based solely on the fear of irreversible cardiotoxicity.
Initial data suggested that the incidence of substantial cardiac dysfunction with trastuzumab may be as high as 16% in patients receiving concomitant doxorubicin and cyclophosphamide,15 but later reports have suggested a much lower incidence.21 This difference may be related to whether or not the chemotherapeutic agents were used concurrently. The most recent report of epirubicin being used in combination with trastuzumab indicated an even lower incidence of LV dysfunction than the incidence seen with doxorubicin, and there were no episodes of severe LV dysfunction.20 Because many patients are treated with combinations of both an anthracycline and cyclophosphamide, it is difficult to determine what the relative risks of these agents might be had they been used alone. It is presumed that the agent used, the method of anthracycline administration (bolus v longer infusion), and the cumulative dose may greatly affect the incidence of cardiotoxicity.11,12,14
Sequential stress may also contribute to myocardial dysfunction.22 During acute injury, such as with myocardial stunning (postischemic reversible myocardial dysfunction), LV dysfunction may be severe and then subsequently improve.23 Pre-existing damage or other conditions (eg, severe hypertension) may magnify the condition, producing profound LV dysfunction. Furthermore, if coexistent conditions, such as hypertension or ischemic heart disease promote LV dysfunction, then the incidence of chemotherapy-induced cardiotoxicity will be higher. In our study, subclinical myocardial damage may have already occurred in the patients because all but one were treated with anthracyclines before being treated with trastuzumab, and the sequential stress is likely to have been a contributing factor.22
The question of incidence of LV dysfunction is also related to the clinical detection of cardiotoxicity. Currently, clinical evidence suggests that detection is best accomplished via sequential measurements of LV function, either by multiple-gated acquisition scans or echocardiography techniques.17 However, the frequency of testing has not been established even with doxorubicin, which is the agent with the highest incidence of cardiotoxicity. These measurements do not provide early detection of abnormalities, probably, in part, because of physiologic compensations for significant cardiac damage; only after such compensation is no longer possible are abnormalities evident. Additionally, both of these typical measurements of LV function vary considerably and are dependent on loading conditions of the LV. Nevertheless, such testing has been of fundamental importance in large clinical trials that have looked at cardiotoxicity because the individual variation and interpretative artifacts are largely equalized when the results from large numbers of patients are pooled for statistical purposes.
Data suggest that other screening tests could reveal cardiotoxicity from chemotherapeutic agents in a more sensitive and immediate manner, such as the measurement of cardiac troponin I. Cardiac troponin, a component of the myocardial contractile apparatus, is released after just one cycle of chemotherapy and is more frequently abnormal with each cycle in patients who ultimately developed LV systolic dysfunction.24-26 b-type natriuretic peptide has also shown limited promise as an immediate marker for damage in patients who subsequently develop CHF.27,28 These markers are potentially advantageous when compared with measuring EF because both can be used during chemotherapy administration and provide immediate feedback regarding risk of subsequent LV dysfunction and CHF. However, their clinical utility has yet to be established.
Cardiac biopsy may be the most sensitive test for detecting anthracycline cardiac damage. Vacuolization and myofibrillar dropout may be detected with electron microscopic evaluation of specimens even in patients who have received as little as 100 to 200 mg/m2 of doxorubicin.29 However, the biopsy abnormalities reflect only recent and ongoing damage, and do not reflect damage that occurred in the past. In a group of six breast cancer patients who had received 300 mg/m2 of doxorubicin in the adjuvant setting and who underwent cardiac biopsy more than 8 months after completion of their adjuvant regimen, all of the patients demonstrated a biopsy grade of 0, suggesting no identifiable toxicity. The biopsy grades, had the material been obtained at the usual biopsy time of 2 to 3 weeks after the administration of doxorubicin, would have been expected to be 0.5 to 1.5.30 Although the hearts of previously treated patients remain functionally abnormal because of their prior exposure and although these abnormalities become clearly evident should these patients be rechallenged with anthracyclines, they no longer manifest these abnormalities through biopsy changes. The fact that the patients in our present review did not exhibit biopsy changes suggests that the anthracycline damage detectable on biopsy was no longer evident and that the dysfunction related to trastuzumab was not associated with the same type of biopsy-detectable damage seen with anthracyclines.31 New techniques need to be developed that can allow the myocardium to be examined for more than just classic ultrastructural abnormalities if myocardial biopsy is to have an important role in detecting nonanthracycline toxicity.
The mechanism of trastuzumab-related cardiotoxicity remains uncertain, but the lack of diagnostic structural findings on myocardial biopsy samples clearly suggests that it differs from the mechanism of anthracyclines. The prompt reversibility of LV dysfunction seen in our patients on discontinuing trastuzumab also supports this notion. A likely mechanism of trastuzumab-related LV dysfunction may involve the signaling pathways that are interrupted by the monoclonal antibody, which, when removed, allows the signaling pathways to recover. Data from animal studies indicate that signaling through an HER2 pathway is necessary to prevent dilated cardiomyopathy.32 It is unknown whether or not this is the same in humans, but perhaps myocardial sampling from biopsies can provide a basis for detailed investigation.
In summary, the mechanism for trastuzumab-related cardiotoxicity is different from that of anthracyclines.31 The clinical outcome is more favorable. The trastuzumab-related cardiotoxicity in patients with HER2/neu-positive breast cancer seems to be largely reversible when trastuzumab is withdrawn. Standard therapy for LV dysfunction and CHF may hasten this recovery; furthermore, such therapy may be cardioprotective, allowing trastuzumab to be reintroduced to patients who have shown a compelling oncologic response to the treatment. A small percentage of such patients had recurrence of LV dysfunction for which further trastuzumab therapy would be contraindicated. This study provides some important clinical evidence that may assist clinicians in decision making regarding trastuzumab-related cardiotoxicity.
Authors’ Disclosures of Potential Conflicts of Interest
The author or 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 discription 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.
| Authors | Employment | Leadership | Consultant | Stock | Honoraria | Research Funds | Testimony | Other |
|---|---|---|---|---|---|---|---|---|
| Michael S. Ewer | Consultant for Bohringer Ingelheim (A) | Pfizer, Speakers’ Bureau (A) |
Dollar Amount Codes (A) < $10,000 (B) $10,000–99,000 (C) ≥ $100,000 (N/R) Not Required
Changes in left ventricular ejection fraction (LVEF) from baseline to re-treatment with trastuzumab in a selected population. (*) P < .05 versus before trastuzumab therapy. (+) P < .05 versus after trastuzumab therapy. (†) Prior anthracyclines in 37 of 38 patients.
Change in left ventricular ejection fraction (LVEF) from baseline to rechallenge with trastuzumab in individual patients.
Clinical algorithm for patients with trastuzumab-related left ventricular (LV) dysfunction. (*) Patients with symptomatic congestive heart failure (CHF). T, trastuzumab; Rx for CHF, standard therapy for congestive heart failure; LVEF, left ventricular ejection fraction.
Typical biopsy (×4,000) from a patient with left ventricular dysfunction on trastuzumab alone. Representative biopsy specimen from a patient with trastuzumab-related cardiotoxicity demonstrating no ultrastructural changes.
Identified Pre-Existing Risk Factors and Prior Chemotherapy Exposure in Patients With Trastuzumab-Related Left Ventricular Dysfunction
Acknowledgments
We thank Amy Chiu for the expert preparation of this manuscript and Rachel Williams for editorial assistance.
Footnotes
-
Presented in part at the 8th Annual Scientific Meeting of the heart failure Society of America, Toronto, Ontario, Canada, September 9-15, 2004.
Authors’ disclosures of potential conflicts of interest are found at the end of this article.
- Received January 24, 2005.
- Accepted May 27, 2005.
