Chemotherapy and Cardiotoxicity in Older Breast Cancer Patients: A Population-Based Study
http://www.100md.com
《临床肿瘤学》
the Department of Medicine and the Herbert Irving Comprehensive Cancer Center, College of Physicians and Surgeons, and Departments of Epidemiology and Biostatistics, Mailman School of Public Health, Columbia University
New York Presbyterian Hospital, New York, NY
ABSTRACT
PURPOSE: Adjuvant chemotherapy, especially with anthracyclines, is known to cause acute and chronic cardiotoxicity in breast cancer patients. We studied the cardiac effects of chemotherapy in a population-based sample of breast cancer patients aged 65 years with long-term follow-up.
PATIENTS AND METHODS: In the Surveillance, Epidemiology, and End Results (SEER)-Medicare database, we analyzed treatments and outcomes among women 65 years of age who were diagnosed with stage I to III breast cancer from January 1, 1992 to December 31, 1999. Propensity scores were used to control for baseline heart disease (HD) and other known predictors of chemotherapy, and Cox proportional hazards models were used to estimate the risk of cardiomyopathy (CM), congestive heart failure (CHF), and HD after chemotherapy.
RESULTS: Of 31,748 women with stage I to III breast cancer, 5,575 (18%) received chemotherapy. Chemotherapy was associated with younger age, fewer comorbidities, hormone receptor negativity, multiple primary tumors, and advanced disease. Patients who received chemotherapy were less likely than other patients to have pre-existing HD (45% v 55%, respectively; P < .001). The hazard ratios for CM, CHF, and HD for patients treated with doxorubicin (DOX) compared with patients who received no chemotherapy were 2.48 (95% CI, 2.10 to 2.93), 1.38 (95% CI, 1.25 to 1.52), and 1.35 (95% CI, 1.26 to 1.44), respectively. The relative risk of cardiotoxicity among patients who received DOX compared with untreated patients remained elevated 5 years after diagnosis.
CONCLUSION: When baseline HD was taken into account, chemotherapy, especially with anthracyclines, was associated with a substantially increased risk of CM. As the number of long-term survivors grows, identifying and minimizing the late effects of treatment will become increasingly important.
INTRODUCTION
Anthracycline-containing chemotherapy improves disease-free and overall survival in patients with breast cancer,1,2 but it is also cardiotoxic.3-6Anthracyclines induce cardiomyopathy (CM), as Lefrak et al7 first described 30 years ago, by multiple mechanisms of myocardial cellular disruption and free radical formation.8,9 For more than 25 years, doxorubicin (DOX) has been the most commonly used anthracycline in the United States, but its long-term sequelae in breast cancer patients have received little attention because DOX was originally reserved for patients with metastatic disease and poor prospects of survival. Its effects on patients with pre-existing cardiac risk factors are also undocumented because such patients are usually excluded from participating in clinical trials.10
In the United States, nearly 50% of patients diagnosed with breast cancer are women greater than 65 years old.11 Although adjuvant chemotherapy has been shown to prolong survival in breast cancer patients,12 its risk-benefit profile in the elderly is poorly understood because elderly breast cancer patients are notably underrepresented in clinical trials.13 Observational studies have demonstrated that breast cancer mortality increases with age14 and that age is a barrier to receiving adjuvant chemotherapy.15-22 Physician and patient concerns about treatment toxicity may have limited chemotherapy use in elderly women for good reason.23 Elderly female breast cancer patients may be at a particularly high risk of treatment-related cardiotoxicity,24 and women have been found to be more susceptible than men to anthracycline-induced cardiotoxicity.25-28
In the year after a course of anthracycline therapy, some patients develop abnormal left ventricular function that may progress to congestive heart failure (CHF).29-32 Estimates of the incidence of anthracycline-induced CHF diagnosed within a year of treatment range from less than 5% to greater than 50% for cumulative doses of 550 to 1,000 mg/m2, respectively.5,33,34 Recently, Perez et al35 found that asymptomatic acute anthracycline-induced cardiotoxicity occurred after only four cycles of anthracycline-containing chemotherapy (240 mg/m2). In addition to acute cardiotoxicity, which is typically transient and clinically manageable, asymptomatic cardiotoxicity may progress to symptomatic CM. Chronic toxicity results from progressive loss of cardiac myocytes. Chronic anthracycline-induced cardiotoxicity may manifest after a prolonged asymptomatic interval as ventricular dysfunction, heart failure, conduction disturbances, and arrhythmias, occurring even decades after the discontinuation of anthracycline therapy.36-38
In the present study, our objective was to evaluate chemotherapy use and cardiotoxicity among elderly women in the general population diagnosed with early-stage breast cancer. This study is unique in taking into account pre-existing cardiac comorbidities and evaluating long-term cardiac risks. We hypothesized that elderly breast cancer patients who received chemotherapy, especially DOX-based chemotherapy, would experience higher rates of cardiotoxicity than patients who did not receive chemotherapy.
PATIENTS AND METHODS
Study Database
This study used a database codeveloped by the US National Cancer Institute and the Center for Medicare and Medicaid Services. The Surveillance, Epidemiology, and End-Results (SEER) program, sponsored by the National Cancer Institute, is a network of tumor registries covering approximately 14% of the US population. The Center for Medicare and Medicaid Services–sponsored Medicare program covers hospital services, physician services, some drug therapy, and other medical services for more than 97% of persons aged 65 years and older. The linked SEER-Medicare database contains clinical, demographic, and medical claims data on patients greater than 65 years of age diagnosed with cancer since 1990. The SEER-Medicare data represent a unique population-based longitudinal data set available for epidemiologic evaluation that has been described comprehensively elsewhere.39
Patient Selection Criteria
We conducted a retrospective cohort study of women, aged 65 years and older and participating in Medicare, who were diagnosed with breast cancer from 1992 to 1999. We excluded women who were diagnosed in 1991 because the database did not include Medicare data on these patients for the year before their diagnosis; women who participated in a Health Maintenance Organization during any month of the study period because data were unavailable for these periods; women who did not participate in both Medicare Part A and Part B during any month of the study period because data were partially unavailable for these periods; women diagnosed with American Joint Committee on Cancer (AJCC) stage 0 or stage IV disease because our evaluation was focused on adjuvant treatment; and women who had a prior cancer diagnosed before age 65 years, a prior breast cancer or other cancer, end-stage renal disease, and a diagnosis without histologic confirmation. For each patient, information was collected from 12 months before breast cancer diagnosis to death or censoring on December 31, 2001.
Measurement of Treatments and Outcomes
Surgery, radiation therapy (RT), and chemotherapy exposures were defined according to previously published criteria using the SEER-Medicare linked databases. Others have found that International Classification of Disease, ninth revision, clinical modification (ICD-9-CM) procedure, Current Procedural Technology (CPT-4), and Health Care Finance Administration Common Procedure Coding System (HCPCS) codes capture 99% of breast cancer patients, with revenue center and ICD-9-CM V codes capturing the balance.22 Chemotherapy agents were found in level II HCPCS codes.40
Mastectomy exposure included total, subcutaneous, radical, or modified radical mastectomies. Breast-conserving surgery exposure included segmental mastectomy, lumpectomy, quadrantectomy, tylectomy, wedge resection, excisional biopsy, nipple resection, and partial mastectomy unspecified. RT exposure included beam RT, radioactive implants, radioisotopes, or other RT documented by SEER but excluded RT planning, hyperthermia, and nuclear medicine codes.41 Chemotherapy exposure was ascertained from the Medicare files using codes for ICD-9-CM diagnosis, ICD-9-CM procedural, CPT, HCPCS, and revenue centers. Among the chemotherapy patients, we created a subgroup of patients who received any DOX-containing regimen. To facilitate comparisons with clinical practice and randomized trials, we also created the following two subgroups of patients who received chemotherapy: patients who received the combination of cyclophosphamide, DOX, and fluorouracil (CAF; a subset of the DOX group) and patients who received the combination of cyclophosphamide, methotrexate, and fluorouracil (CMF).
We developed a coding algorithm using diagnosis and procedure codes in the Medicare files to identify cardiotoxicity outcomes. We measured the following outcomes: CM, CHF, heart disease (HD), and myocardial infarction (MI). The composite measure HD included diagnosis codes for CM, acute myocarditis, CHF, acute MI, arrhythmias, ventricular dysfunction, ischemic cardiac disease, and sudden death.
Comorbidity and Propensity Scoring
We used a comorbidity index, which was developed by Klabunde et al42 and based on conditions identified by Charlson et al,43 to search for diagnosis and procedure claims in the inpatient and outpatient Medicare data. The Charlson scale is considered to be a reliable measure of comorbidity in cancer trials of older patients44 and has been found to be a predictive measure of hospitalization in breast cancer patients using the SEER-Medicare linked database.45 As a corollary to this measurement, we applied our cardiac outcome-seeking algorithm to the year before breast cancer diagnosis to identify patients with pre-existing HD. We used the comorbidity score, the pre-existing HD variable, and demographic and clinical factors to control for baseline differences between groups in our multivariate analysis.
Propensity scoring is a technique used to measure the likelihood that a patient will be treated solely based on that patient's independent predictors of treatment.46,47 We used logistic regression to evaluate the odds of receiving chemotherapy among breast cancer patients according to baseline risk factors. We then used the coefficients for the predictors to assign to each patient a composite propensity score that could be used in lieu of the covariates included in the multivariate analysis. The results derived from using the propensity score in lieu of the covariates did not differ appreciably from the multivariate analysis; therefore, the multivariate analysis is presented.
Statistical Analysis
The 2 test and analysis of variance were used to compare frequency distributions of categoric and continuous variables, respectively. All hypothesis tests were two sided. We used stratified analyses to test the main effect in treatment subgroups, control for confounding, and describe effect modification.
We performed logistic and Cox proportional hazards regression. The logistic regression modeled the log odds of experiencing a cardiac outcome while controlling for covariates, such as year of diagnosis, age, race, stage, tumor histology, node status, estrogen receptor status, and comorbidities. We used the Hosmer-Lemeshow goodness-of-fit test to compare larger models to smaller models so that the most parsimonious model was selected to represent the data. The results are described in terms of odds ratios and 95% CIs. Using Cox proportional hazards models, we developed rate ratios of cardiac events comparing treated with untreated patients. Separate models were created for any chemotherapy, CMF, DOX, and CAF with each of the outcome variables (HD, CHF, CM, and MI). The comparison group in each of these analyses is patients who did not receive chemotherapy. Maximum partial likelihood estimates of hazard ratios with 95% CIs were obtained. We generated Kaplan-Meier survival curves and applied the log-rank test to compare cardiac disease-free survival across groups. All statistical analyses were conducted using the SAS system for Windows version 8.02 (SAS Institute, Cary, NC).
RESULTS
The SEER-Medicare database included 152,200 women aged 65 years and older who were diagnosed with breast cancer from 1992 to 1999. We excluded 40,170 women who were diagnosed from 1991 to 1992; 30,430 women who participated in a Health Maintenance Organization during any month of the study period; 33,611 women who did not participate in both Medicare Part A and Part B during any month of the study period; 5,339 women with AJCC stage 0 and 5,114 women with AJCC stage IV disease; and 5,788 women for one or more of the other reasons for exclusion.
The final study cohort consisted of 31,748 women with histologically confirmed AJCC stage I to III breast cancer diagnosed between the years 1992 and 2000. Of these women, 5,575 (17.6%) received chemotherapy. Patients who received chemotherapy, compared with patients who did not receive chemotherapy, were younger (mean age, 72.5 v 76.2 years, respectively; P < .0001), diagnosed at a later stage, more likely to be hormone receptor negative, more likely to undergo mastectomy and RT, and less likely to have pre-existing HD (36% v 45%, respectively; P < .001) or other comorbidities at baseline (Table 1).
Among patients who received chemotherapy, the likelihood of receiving DOX increased from 21% in 1992 to 48% in 1999. Chemotherapy patients who received DOX, compared with untreated patients, were younger (71 v 73 years, respectively; P < .0001), more likely to be diagnosed with later stage disease, and more likely to have hormone receptor–negative disease (24% v 7%, respectively; P < .001); DOX patients were also less likely to have pre-existing HD (29% v 45%, respectively; P < .001) or any other comorbidities at baseline (Table 1).
Chemotherapy use decreased with increasing age. Compared with patients who were 65 to 69 years old, patients who were 70 to 74, 75 to 79, and more than 80 years old were 38%, 62%, and 85% less likely to receive chemotherapy, respectively.
The propensity to receive any chemotherapy was also associated with number of comorbidities, hormone receptor negativity, mastectomy, and RT. The odds of receiving DOX increased over time, and the odds of receiving CAF decreased over time, reflecting a possible increase in the use of DOX plus cyclophosphamide. Notably, patients with pre-existing HD were less likely than other patients to receive DOX (Table 2).
Cox proportional hazards models evaluating chemotherapy and HD are shown in Table 3. Increasing age was associated with an increased rate of cardiac diseases during the study period. Compared with white patients, black patients had a 25% higher rate of HD. Patients diagnosed with HD during the year before breast cancer diagnosis were 2.4 times more likely to experience HD during follow-up. Controlling for these covariates, patients receiving chemotherapy had a 22% increase in the rate of HD, a 55% increase in the rate of CM diagnosis, and a 20% increase in the rate of CHF diagnosis compared with untreated patients. In a separate analysis of women receiving both RT and chemotherapy, the combination was associated with a slightly higher risk of HD (33%).
We also evaluated interaction terms in the models to determine what factors contributed to an increased risk of adverse cardiac outcome. We found that black patients are at highest risk of chemotherapy-induced HD (P < .007), and patients with pre-existing HD (P < .0007) and two or more comorbidities (P = .05) are at highest risk for DOX-induced CM. The interaction term in the model for RT and chemotherapy and the risk of HD was not significant (P = .84).
Patients treated with DOX were 2.5 times more likely to experience CM than untreated patients. Patients treated with CAF had higher rates of all cardiac outcomes than patients treated with CMF. The difference in hazard ratios between the two regimens was most pronounced for CM. Compared with untreated patients, CAF patients were 2.3 times more likely to experience CM, whereas CMF patients were only 1.3 times more likely to experience CM (Table 4).
The relationship between DOX use and subsequent CM was the most profound association that we found. We tested the proportionality of the hazards and found that the relationship between CM and DOX was independent of time. Chemotherapy use was not associated with an increased risk of MI. Even when using a conservative adjustment for multiple statistical tests, the P values for all of the estimates were less than the adjusted P value threshold of .007.
In the first year after diagnosis, the incidence of CM ranged from a low of 1.55% for patients not receiving chemotherapy to 4.09% for patients receiving DOX (odds ratio = 3.51; 95% CI, 2.63 to 4.69). After 5 years of follow-up, the cumulative incidence had increased to 4.97% in patients not receiving chemotherapy and 10.23% in patients receiving CAF. Figure 1A presents the annual incidence of CM by year since breast cancer diagnosis. Figure 1B presents these data cumulatively by year since diagnosis.
CM-free survival was defined as time to first CM diagnosis during the follow-up period. Figure 2 presents the Kaplan-Meier curves for CM-free survival for patients receiving CMF, DOX, and CAF compared with patients who did not receive chemotherapy. There was no difference between the Kaplan-Meier curve for patients who received CMF and the curve for patients who did not receive chemotherapy. The Kaplan-Meier curve for patients not receiving chemotherapy was significantly different from the curve for patients receiving DOX and CAF according to the log-rank test (P < .0001).
DISCUSSION
Our analysis showed that elderly women with breast cancer who had pre-existing cardiac disease were less likely than women without such disease to be treated with chemotherapy. We also found that, despite being selected for favorable cardiac risk, patients who received chemotherapy were more likely than patients who did not receive chemotherapy to experience cardiac disease subsequently. DOX use was associated with a three-fold increase in the rate of CM in the first year after treatment. As in prior studies, we found that the incidence and severity of anthracycline-induced cardiotoxicity seems to persist over time.22,51,52 Although we observed that the amplified risk of treatment-related CM decreased considerably after the first year, the risk of experiencing CM for DOX patients was still 50% higher than the risk for patients who did not receive chemotherapy in the fifth year after diagnosis.
Anthracycline-induced CM has an overall incidence of 1% to 5% that varies with the cumulative dose of anthracycline.9,48 Previous research suggests that DOX-induced cardiotoxicity may be more pronounced in the elderly. Patients more than 65 years old are 2.25 times more likely to experience CHF compared with patients less than 65 years old.49 Recently, Perez et al35 found that 23% of patients treated with four cycles of DOX plus cyclophosphamide experienced toxicity immediately after treatment, suggesting that the damage may be frequent yet reversible. Anthracycline-induced cardiotoxicity is also related to the rate of administration,50 sex,26 pre-existing HD,5 and left-sided RT.24,51,52
Our data did not include either cumulative dose of DOX or the method of administration. However, because standard treatment usually consists of doses between 240 and 360 mg/m2, it is unlikely that any patient received more than 400 mg/m2. In a prior study, we found that older patients were more likely than younger patients to terminate treatment early.53 Therefore, we consider it likely that our patients received relatively low cumulative doses. If so, our estimates of the cardiac effects of DOX should be viewed as conservative. Although our study does not provide insight into the debate on threshold dose and optimal schedule for anthracycline-induced cardiotoxicity, it does capture baseline risk factors and real-world utilization of chemotherapy in a population-based cohort of elderly breast cancer patients suffering comorbidities.
Previous studies have suggested that RT is an independent risk factor for anthracycline-induced cardiotoxicity.54-56 Many women are treated with a combination of RT and chemotherapy; yet few long-term data are available on the interactive effects of these treatments.57,58 RT did not independently predict cardiac outcomes after controlling for other covariates; however, RT does seem to potentiate the effect of chemotherapy. The results of our investigation of RT and cardiotoxicity are reported separately (Doyle et al, manuscript in preparation).
A few studies have reported no increase in the risk of cardiotoxicity after anthracycline chemotherapy. Among patients treated in a clinical trial, Ganz et al found no difference in cardiac events 5 to 8 years after treatment. The findings from this study and other clinical trials have limited generalizability to our elderly study population because the studies typically exclude patients who are older than 65 years and who are diagnosed with pre-existing cardiovascular disease.59 We found that patients who received CMF also had higher risk of cardiotoxicity than patients who did not receive chemotherapy.
Others have found that older age and comorbidities are associated with less aggressive treatment in early-stage breast cancer.22,60 Indeed, our study demonstrates that patients who were older, diagnosed at an earlier stage, hormone receptor positive, or suffering from pre-existing HD or any comorbidities were less likely to receive chemotherapy than other patients. Newcomb and Carbone17 found that only 35% of patients older than 65 years were offered chemotherapy and that such patients were twice as likely as younger patients to reject chemotherapy for fear of side effects. However, data from clinical trials suggest that older women and younger women derived similar reductions in breast cancer mortality from chemotherapy, and age alone should not be a contraindication to the use of chemotherapy in older women who are in good health.61
Selection of patients for chemotherapy was undoubtedly influenced by cardiac risk. Therefore, the chemotherapy patient cohort consists of fewer patients with cardiac disease. Both the Charlson-derived comorbidity score and pre-existing HD diagnosis were associated with subsequent cardiac events. These findings are consistent with other studies that have shown that breast cancer patients suffering from comorbidities experience worse outcomes than breast cancer patients without comorbidities.62,63
The SEER-Medicare database is an invaluable tool for studying unanticipated treatment effects and long-term outcomes in a population-based sample of patients who, for various reasons, have been underrepresented in clinical trials. Given the constraints on eligibility for such trials and other barriers to trial participation, using such databases is the only way to find out how treatments work in the real world. Our study extends the findings of clinical trial research conducted among younger women to an elderly breast cancer population.
The observational design of the study and incomplete measurement of risk factors related to HD impose some limitations on the interpretation of our findings. We used propensity score analysis to control for potential confounders on the association between treatment and outcomes. Inadequate control for health-related factors may have led to an underestimate of the cardiac risks associated with chemotherapy. However, even within the selected group of patients who received chemotherapy, patients who were treated with DOX-based chemotherapy had a higher risk of cardiotoxicity than patients who received other regimens. We doubt that this association is a result of selection bias. We also believe that even a small increase in the risk of a common and serious adverse health effect, such as cardiac disease, has important public health implications.
Our study demonstrates that the elevated risk of cardiac disease associated with adjuvant chemotherapy, which has been observed among participants in clinical trials, is also present among elderly breast cancer patients in a population-based sample. Within that sample, this risk is associated with older age, is higher among patients who receive anthracyclines, and, although most pronounced in the first year after treatment, continues over time. Given the prevalence of anthracycline regimens in the current adjuvant breast cancer setting, some patients may benefit from increased cardiac monitoring. Because the effects of chemotherapy may appear late in long-term survivors of breast cancer, there is a need for new models to assess the long-term effects of treatment.
Authors' Disclosures of Potential Conflicts of Interest
The authors indicated no potential conflicts of interest.
Acknowledgment
We acknowledge the efforts of the Applied Research Branch, Division of Cancer Prevention and Population Science, National Cancer Institute; the Office of Information Services and the Office of Strategic Planning, Health Care Financing Administration; Information Management Services, Inc; and the Surveillance, Epidemiology, and End Results (SEER) Program tumor registries in the creation of the SEER-Medicare database.
NOTES
Supported by an American Society of Clinical Oncology Career Development Award (D.L.H.), K07 Award No. CA95597 from the National Cancer Institute (D.L.H.), K05 Award No. CA89155 from the National Cancer Institute (A.I.N.), and Grant No. RSGT-01-024-04-CPHPS from the American Cancer Society (A.I.N.).
This study used the linked Surveillance, Epidemiology, and End Results-Medicare database. The interpretation and reporting of these data are the sole responsibility of the authors.
Authors' disclosures of potential conflicts of interest are found at the end of this article.
REFERENCES
Levine MN, Bramwell VH, Pritchard KI, et al: Randomized trial of intensive cyclophosphamide, epirubicin, and fluorouracil chemotherapy compared with cyclophosphamide, methotrexate, and fluorouracil in premenopausal women with node-positive breast cancer: National Cancer Institute of Canada Clinical Trials Group. J Clin Oncol 16: 2651-2658, 1998
Basser L, Abraham R, To LB, et al: Cardiac effects of high-dose epirubicin and cyclophosphamide in women with poor prognosis breast cancer. Ann Oncol 10: 53-58, 1999
Sparano JA: Use of dexrazoxane and other strategies to prevent cardiomyopathy associated with doxorubicin-taxane combinations. Semin Oncol 25: 66-71, 1998
Singal PK, Iliskovic N: Doxorubicin-induced cardiomyopathy. N Engl J Med 339: 900-905, 1998
von Hoff DD, Layard MW, Basa P, et al: Risk factors for doxorubicin-induced congestive heart failure. Ann Intern Med 91: 710-717, 1979
Zhou S, Starkov A, Froberg MK, et al: Cumulative and irreversible cardiac mitochondrial dysfunction induced by doxorubicin. Cancer Res 61: 771-777, 2001
Lefrak EA, Pitha J, Rosenheim S, et al: A clinicopathologic analysis of Adriamycin cardiotoxicity. Cancer 32: 302-314, 1973
Gewirtz DA: A critical evaluation of the mechanisms of action proposed for the antitumor effects of the anthracycline antibiotics adriamycin and daunorubicin. Biochem Pharmacol 57: 727-741, 1999
Hortobagyi GN: Anthracyclines in the treatment of cancer: An overview. Drugs 54: 1-7, 1997
Zambetti M, Moliterni A, Materazzo C, et al: Long-term cardiac sequelae in operable breast cancer patients given adjuvant chemotherapy with or without doxorubicin and breast irradiation. J Clin Oncol 19: 37-43, 2001
Crivellari D, Bonetti M, Castiglione-Gertsch M, et al: Burdens and benefits of adjuvant cyclophosphamide, methotrexate, and fluorouracil and tamoxifen for elderly patients with breast cancer: The International Breast Cancer Study Group Trial VII. J Clin Oncol 18: 1412-1422, 2000
Fumoleau P, Kerbrat P, Romestaing P, et al: Randomized trial comparing six versus three cycles of epirubicin-based adjuvant chemotherapy in premenopausal, node-positive breast cancer patients: 10-Year follow-up results of the French Adjuvant Study Group 01 Trial. J Clin Oncol 21: 298-305, 2003
Hutchins LF, Unger JM, Crowley JJ, et al: Underrepresentation of patients 65 years of age or older in cancer-treatment trials. N Engl J Med 341: 2061-2067, 1999
Extermann M, Balducci L, Lyman GH: What threshold for adjuvant therapy in older breast cancer patients? J Clin Oncol 18: 1709-1717, 2000
Hillner BE, Penberthy L, Desch CE, et al: Variation in staging and treatment of local and regional breast cancer in the elderly. Breast Cancer Res Treat 40: 75-86, 1996
Mandelblatt JS, Hadley J, Kerner JF, et al: Patterns of breast carcinoma treatment in older women: Patient preference and clinical and physical influences. Cancer 89: 561-573, 2000
Newcomb PA, Carbone PP: Cancer treatment and age: Patient perspectives. J Natl Cancer Inst 85: 1580-1584, 1993
Chu J, Diehr P, Feigl P, et al: The effect of age on the care of women with breast cancer in community hospitals. J Gerontol 42: 185-190, 1987
Osteen RT, Cady B, Chmiel JS, et al: 1991 national survey of carcinoma of the breast by the Commission on Cancer. J Am Coll Surg 178: 213-219, 1994
Harlan LC, Abrams J, Warren JL, et al: Adjuvant therapy for breast cancer: Practice patterns of community physicians. J Clin Oncol 20: 1809-1817, 2002
Du XL, Key CR, Osborne C, et al: Discrepancy between consensus recommendations and actual community use of adjuvant chemotherapy in women with breast cancer. Ann Intern Med 138: 90-97, 2003
Du X, Goodwin JS: Patterns of use of chemotherapy for breast cancer in older women: Findings from Medicare claims data. J Clin Oncol 19: 1455-1461, 2001
Hebert-Croteau N, Brisson J, Latreille J, et al: Time trends in systemic adjuvant treatment for node-negative breast cancer. J Clin Oncol 17: 1458-1464, 1999
Bristow MR, Billingham ME, Mason JW, et al: Clinical spectrum of anthracycline antibiotic cardiotoxicity. Cancer Treat Rep 62: 873-879, 1978
Hrushesky WJ, Fader DJ, Berestka JS, et al: Diminishment of respiratory sinus arrhythmia foreshadows doxorubicin-induced cardiomyopathy. Circulation 84: 697-707, 1991
Lipshultz SE, Lipsitz SR, Mone SM, et al: Female sex and drug dose as risk factors for late cardiotoxic effects of doxorubicin therapy for childhood cancer. N Engl J Med 332: 1738-1743, 1995
Silber JH, Barber G: Doxorubicin-induced cardiotoxicity. N Engl J Med 333: 1359-1360, 1995
Krischer J, Epstein S, Cuthbertson D, et al: Clinical cardiotoxicity following anthracycline treatment for childhood cancer: The Pediatric Oncology Group experience. J Clin Oncol 15: 1544-1552, 1997
Bristow MR, Thompson PD, Martin RP, et al: Early anthracycline cardiotoxicity. Am J Med 65: 823-832, 1978
Shan K, Lincoff AM, Young JB: Anthracycline-induced cardiotoxicity. Ann Intern Med 125: 47-58, 1996
Grenier MA, Lipshultz SE: Epidemiology of anthracycline cardiotoxicity in children and adults. Semin Oncol 25: 72-85, 1998
Steinberg JS, Cohen AJ, Wasserman AG, et al: Acute arrhythmogenicity of doxorubicin administration. Cancer 60: 1213-1218, 1987
von Hoff DD, Rozencweig M, Piccart M: The cardiotoxicity of anticancer agents. Semin Oncol 9: 23-33, 1982
Allen A: The cardiotoxicity of chemotherapeutic drugs. Semin Oncol 19: 529-542, 1992
Perez EA, Suman VJ, Davidson NE, et al: Effect of doxorubicin plus cyclophosphamide on left ventricular ejection fraction in patients with breast cancer in the North Central Cancer Treatment Group N9831 Intergroup Adjuvant Trial. J Clin Oncol 22: 3700-3704, 2004
Freter CE, Lee TC, Billingham ME, et al: Doxorubicin cardiac toxicity manifesting seven years after treatment: Case report and review. Am J Med 80: 483-485, 1986
Steinherz LJ, Steinherz PG, Tan CT, et al: Cardiac toxicity 4 to 20 years after completing anthracycline therapy. JAMA 266: 1672-1677, 1991
Gottlieb SL, Edmiston WA Jr, Haywood LJ: Late, late doxorubicin cardiotoxicity. Chest 78: 880-882, 1980
Warren JL, Klabunde CN, Schrag D, et al: Overview of the SEER-Medicare data: Content, research applications, and generalizability to the United States elderly population. Med Care 40: IV–3-IV–18, 2002 (suppl 8)
Warren JL, Harlan LC, Fahey A, et al: Utility of the SEER-Medicare data to identify chemotherapy use. Med Care 40: IV–55-IV–61, 2002 (suppl 8)
Du X, Freeman JL, Goodwin JS: Information on radiation treatment in patients with breast cancer: The advantages of the linked medicare and SEER data—Surveillance, Epidemiology and End Results. J Clin Epidemiol 52: 463-470, 1999
Klabunde CN, Potosky AL, Legler JM, et al: Development of a comorbidity index using physician claims data. J Clin Epidemiol 53: 1258-1267, 2000
Charlson ME, Pompei P, Ales KL, et al: A new method of classifying prognostic comorbidity in longitudinal studies: Development and validation. J Chronic Dis 40: 373-383, 1986
Extermann M, Overcash J, Lyman GH, et al: Comorbidity and functional status are independent in older cancer patients. J Clin Oncol 16: 1582-1587, 1998
Du XL, Osborne C, Goodwin JS: Population-based assessment of hospitalizations for toxicity from chemotherapy in older women with breast cancer. J Clin Oncol 20: 4636-4642, 2002
D'Agostino RB: Tutorial in biostatistics: Propensity score methods for bias reduction in the comparison of a treatment to a non-randomized control group. Stat Med 17: 2265-2281, 1998
Rubin DB: Estimating causal effects from large data sets using propensity scores. Ann Intern Med 127: 757-763, 1997
Hahm S, Dresner HS, Podwall D, et al: DNA biomarkers antecede semiquantitative anthracycline cardiomyopathy. Cancer Invest 21: 53-67, 2003
Swain SM, Whaley FS, Ewer MS: Congestive heart failure in patients treated with doxorubicin: A retrospective analysis of three trials. Cancer 97: 2869-2879, 2003
Legha SS, Benjamin RS, Mackay B, et al: Reduction of doxorubicin cardiotoxicity by prolonged continuous intravenous infusion. Ann Intern Med 96: 133-139, 1982
Praga C, Beretta G, Vigo PL, et al: Adriamycin cardiotoxicity: A survey of 1273 patients. Cancer Treat Rep 63: 827-834, 1979
Pihkala J, Saarinen UM, Lundstrom U, et al: Myocardial function in children and adolescents after therapy with anthracyclines and chest irradiation. Eur J Cancer 32A: 97-103, 1996
Hershman D, McBride R, Jacobson J, et al: Racial disparities in treatment and survival among women with early-stage breast cancer. J Clin Oncol 23: 6639-6646, 2005
Eltringham RJ, Fajardo L, Stewart RJ, et al: Investigation of cardiotoxicity in rabbits from Adriamycin and fractionated cardiac irradiation preliminary result. Front Radiat Ther Oncol 13: 21-35, 1979
Ryberg M, Nielsen D, Skovsgaard T, et al: Epirubicin cardiotoxicity: An analysis of 469 patients with metastatic breast cancer. J Clin Oncol 16: 3502-3508, 1998
Hardenbergh PH, Recht A, Gollamudi S, et al: Treatment-related toxicity from a randomized trial of the sequencing of doxorubicin and radiation therapy in patients treated for early stage breast cancer. Int J Radiat Oncol Biol Phys 45: 69-72, 1999
Shapiro CL, Recht A: Side effects of adjuvant treatment of breast cancer. N Engl J Med 344: 1997-2008, 2001
Recht A, Come SE, Henderson IC, et al: The sequencing of chemotherapy and radiation therapy after conservative surgery for early-stage breast cancer. N Engl J Med 334: 1356-1361, 1996
Ganz PA, Green SJ, Hutchins L, et al: Late cardiac effects of adjuvant CMF vs CAF in women with node positive breast cancer treated on SWOG 8897: Initial results from SWOG 9342. Presented at the San Antonio Breast Cancer Symposium, San Antonio, TX, December 11-14, 2002
Newschaffer CJ, Penberthy L, Desch CE, et al: The effect of age and comorbidity in the treatment of elderly women with nonmetastatic breast cancer. Arch Intern Med 156: 85-90, 1996
Muss HB, Woolf SH, Cirrincione C, et al: Adjuvant chemotherapy in older and younger women with lymph node-positive breast cancer. JAMA 293: 1073-1081, 2005
Yancik R, Wesley MN, Ries LA, et al: Effect of age and comorbidity in postmenopausal breast cancer patients aged 55 years and older. JAMA 285: 885-892, 2001
Read WL, Tierney RM, Page NC, et al: Differential prognostic impact of comorbidity. J Clin Oncol 22: 3099-3103, 2004
Submitted May 2, 2005; accepted September 13, 2005.(John J. Doyle, Alfred I. )
New York Presbyterian Hospital, New York, NY
ABSTRACT
PURPOSE: Adjuvant chemotherapy, especially with anthracyclines, is known to cause acute and chronic cardiotoxicity in breast cancer patients. We studied the cardiac effects of chemotherapy in a population-based sample of breast cancer patients aged 65 years with long-term follow-up.
PATIENTS AND METHODS: In the Surveillance, Epidemiology, and End Results (SEER)-Medicare database, we analyzed treatments and outcomes among women 65 years of age who were diagnosed with stage I to III breast cancer from January 1, 1992 to December 31, 1999. Propensity scores were used to control for baseline heart disease (HD) and other known predictors of chemotherapy, and Cox proportional hazards models were used to estimate the risk of cardiomyopathy (CM), congestive heart failure (CHF), and HD after chemotherapy.
RESULTS: Of 31,748 women with stage I to III breast cancer, 5,575 (18%) received chemotherapy. Chemotherapy was associated with younger age, fewer comorbidities, hormone receptor negativity, multiple primary tumors, and advanced disease. Patients who received chemotherapy were less likely than other patients to have pre-existing HD (45% v 55%, respectively; P < .001). The hazard ratios for CM, CHF, and HD for patients treated with doxorubicin (DOX) compared with patients who received no chemotherapy were 2.48 (95% CI, 2.10 to 2.93), 1.38 (95% CI, 1.25 to 1.52), and 1.35 (95% CI, 1.26 to 1.44), respectively. The relative risk of cardiotoxicity among patients who received DOX compared with untreated patients remained elevated 5 years after diagnosis.
CONCLUSION: When baseline HD was taken into account, chemotherapy, especially with anthracyclines, was associated with a substantially increased risk of CM. As the number of long-term survivors grows, identifying and minimizing the late effects of treatment will become increasingly important.
INTRODUCTION
Anthracycline-containing chemotherapy improves disease-free and overall survival in patients with breast cancer,1,2 but it is also cardiotoxic.3-6Anthracyclines induce cardiomyopathy (CM), as Lefrak et al7 first described 30 years ago, by multiple mechanisms of myocardial cellular disruption and free radical formation.8,9 For more than 25 years, doxorubicin (DOX) has been the most commonly used anthracycline in the United States, but its long-term sequelae in breast cancer patients have received little attention because DOX was originally reserved for patients with metastatic disease and poor prospects of survival. Its effects on patients with pre-existing cardiac risk factors are also undocumented because such patients are usually excluded from participating in clinical trials.10
In the United States, nearly 50% of patients diagnosed with breast cancer are women greater than 65 years old.11 Although adjuvant chemotherapy has been shown to prolong survival in breast cancer patients,12 its risk-benefit profile in the elderly is poorly understood because elderly breast cancer patients are notably underrepresented in clinical trials.13 Observational studies have demonstrated that breast cancer mortality increases with age14 and that age is a barrier to receiving adjuvant chemotherapy.15-22 Physician and patient concerns about treatment toxicity may have limited chemotherapy use in elderly women for good reason.23 Elderly female breast cancer patients may be at a particularly high risk of treatment-related cardiotoxicity,24 and women have been found to be more susceptible than men to anthracycline-induced cardiotoxicity.25-28
In the year after a course of anthracycline therapy, some patients develop abnormal left ventricular function that may progress to congestive heart failure (CHF).29-32 Estimates of the incidence of anthracycline-induced CHF diagnosed within a year of treatment range from less than 5% to greater than 50% for cumulative doses of 550 to 1,000 mg/m2, respectively.5,33,34 Recently, Perez et al35 found that asymptomatic acute anthracycline-induced cardiotoxicity occurred after only four cycles of anthracycline-containing chemotherapy (240 mg/m2). In addition to acute cardiotoxicity, which is typically transient and clinically manageable, asymptomatic cardiotoxicity may progress to symptomatic CM. Chronic toxicity results from progressive loss of cardiac myocytes. Chronic anthracycline-induced cardiotoxicity may manifest after a prolonged asymptomatic interval as ventricular dysfunction, heart failure, conduction disturbances, and arrhythmias, occurring even decades after the discontinuation of anthracycline therapy.36-38
In the present study, our objective was to evaluate chemotherapy use and cardiotoxicity among elderly women in the general population diagnosed with early-stage breast cancer. This study is unique in taking into account pre-existing cardiac comorbidities and evaluating long-term cardiac risks. We hypothesized that elderly breast cancer patients who received chemotherapy, especially DOX-based chemotherapy, would experience higher rates of cardiotoxicity than patients who did not receive chemotherapy.
PATIENTS AND METHODS
Study Database
This study used a database codeveloped by the US National Cancer Institute and the Center for Medicare and Medicaid Services. The Surveillance, Epidemiology, and End-Results (SEER) program, sponsored by the National Cancer Institute, is a network of tumor registries covering approximately 14% of the US population. The Center for Medicare and Medicaid Services–sponsored Medicare program covers hospital services, physician services, some drug therapy, and other medical services for more than 97% of persons aged 65 years and older. The linked SEER-Medicare database contains clinical, demographic, and medical claims data on patients greater than 65 years of age diagnosed with cancer since 1990. The SEER-Medicare data represent a unique population-based longitudinal data set available for epidemiologic evaluation that has been described comprehensively elsewhere.39
Patient Selection Criteria
We conducted a retrospective cohort study of women, aged 65 years and older and participating in Medicare, who were diagnosed with breast cancer from 1992 to 1999. We excluded women who were diagnosed in 1991 because the database did not include Medicare data on these patients for the year before their diagnosis; women who participated in a Health Maintenance Organization during any month of the study period because data were unavailable for these periods; women who did not participate in both Medicare Part A and Part B during any month of the study period because data were partially unavailable for these periods; women diagnosed with American Joint Committee on Cancer (AJCC) stage 0 or stage IV disease because our evaluation was focused on adjuvant treatment; and women who had a prior cancer diagnosed before age 65 years, a prior breast cancer or other cancer, end-stage renal disease, and a diagnosis without histologic confirmation. For each patient, information was collected from 12 months before breast cancer diagnosis to death or censoring on December 31, 2001.
Measurement of Treatments and Outcomes
Surgery, radiation therapy (RT), and chemotherapy exposures were defined according to previously published criteria using the SEER-Medicare linked databases. Others have found that International Classification of Disease, ninth revision, clinical modification (ICD-9-CM) procedure, Current Procedural Technology (CPT-4), and Health Care Finance Administration Common Procedure Coding System (HCPCS) codes capture 99% of breast cancer patients, with revenue center and ICD-9-CM V codes capturing the balance.22 Chemotherapy agents were found in level II HCPCS codes.40
Mastectomy exposure included total, subcutaneous, radical, or modified radical mastectomies. Breast-conserving surgery exposure included segmental mastectomy, lumpectomy, quadrantectomy, tylectomy, wedge resection, excisional biopsy, nipple resection, and partial mastectomy unspecified. RT exposure included beam RT, radioactive implants, radioisotopes, or other RT documented by SEER but excluded RT planning, hyperthermia, and nuclear medicine codes.41 Chemotherapy exposure was ascertained from the Medicare files using codes for ICD-9-CM diagnosis, ICD-9-CM procedural, CPT, HCPCS, and revenue centers. Among the chemotherapy patients, we created a subgroup of patients who received any DOX-containing regimen. To facilitate comparisons with clinical practice and randomized trials, we also created the following two subgroups of patients who received chemotherapy: patients who received the combination of cyclophosphamide, DOX, and fluorouracil (CAF; a subset of the DOX group) and patients who received the combination of cyclophosphamide, methotrexate, and fluorouracil (CMF).
We developed a coding algorithm using diagnosis and procedure codes in the Medicare files to identify cardiotoxicity outcomes. We measured the following outcomes: CM, CHF, heart disease (HD), and myocardial infarction (MI). The composite measure HD included diagnosis codes for CM, acute myocarditis, CHF, acute MI, arrhythmias, ventricular dysfunction, ischemic cardiac disease, and sudden death.
Comorbidity and Propensity Scoring
We used a comorbidity index, which was developed by Klabunde et al42 and based on conditions identified by Charlson et al,43 to search for diagnosis and procedure claims in the inpatient and outpatient Medicare data. The Charlson scale is considered to be a reliable measure of comorbidity in cancer trials of older patients44 and has been found to be a predictive measure of hospitalization in breast cancer patients using the SEER-Medicare linked database.45 As a corollary to this measurement, we applied our cardiac outcome-seeking algorithm to the year before breast cancer diagnosis to identify patients with pre-existing HD. We used the comorbidity score, the pre-existing HD variable, and demographic and clinical factors to control for baseline differences between groups in our multivariate analysis.
Propensity scoring is a technique used to measure the likelihood that a patient will be treated solely based on that patient's independent predictors of treatment.46,47 We used logistic regression to evaluate the odds of receiving chemotherapy among breast cancer patients according to baseline risk factors. We then used the coefficients for the predictors to assign to each patient a composite propensity score that could be used in lieu of the covariates included in the multivariate analysis. The results derived from using the propensity score in lieu of the covariates did not differ appreciably from the multivariate analysis; therefore, the multivariate analysis is presented.
Statistical Analysis
The 2 test and analysis of variance were used to compare frequency distributions of categoric and continuous variables, respectively. All hypothesis tests were two sided. We used stratified analyses to test the main effect in treatment subgroups, control for confounding, and describe effect modification.
We performed logistic and Cox proportional hazards regression. The logistic regression modeled the log odds of experiencing a cardiac outcome while controlling for covariates, such as year of diagnosis, age, race, stage, tumor histology, node status, estrogen receptor status, and comorbidities. We used the Hosmer-Lemeshow goodness-of-fit test to compare larger models to smaller models so that the most parsimonious model was selected to represent the data. The results are described in terms of odds ratios and 95% CIs. Using Cox proportional hazards models, we developed rate ratios of cardiac events comparing treated with untreated patients. Separate models were created for any chemotherapy, CMF, DOX, and CAF with each of the outcome variables (HD, CHF, CM, and MI). The comparison group in each of these analyses is patients who did not receive chemotherapy. Maximum partial likelihood estimates of hazard ratios with 95% CIs were obtained. We generated Kaplan-Meier survival curves and applied the log-rank test to compare cardiac disease-free survival across groups. All statistical analyses were conducted using the SAS system for Windows version 8.02 (SAS Institute, Cary, NC).
RESULTS
The SEER-Medicare database included 152,200 women aged 65 years and older who were diagnosed with breast cancer from 1992 to 1999. We excluded 40,170 women who were diagnosed from 1991 to 1992; 30,430 women who participated in a Health Maintenance Organization during any month of the study period; 33,611 women who did not participate in both Medicare Part A and Part B during any month of the study period; 5,339 women with AJCC stage 0 and 5,114 women with AJCC stage IV disease; and 5,788 women for one or more of the other reasons for exclusion.
The final study cohort consisted of 31,748 women with histologically confirmed AJCC stage I to III breast cancer diagnosed between the years 1992 and 2000. Of these women, 5,575 (17.6%) received chemotherapy. Patients who received chemotherapy, compared with patients who did not receive chemotherapy, were younger (mean age, 72.5 v 76.2 years, respectively; P < .0001), diagnosed at a later stage, more likely to be hormone receptor negative, more likely to undergo mastectomy and RT, and less likely to have pre-existing HD (36% v 45%, respectively; P < .001) or other comorbidities at baseline (Table 1).
Among patients who received chemotherapy, the likelihood of receiving DOX increased from 21% in 1992 to 48% in 1999. Chemotherapy patients who received DOX, compared with untreated patients, were younger (71 v 73 years, respectively; P < .0001), more likely to be diagnosed with later stage disease, and more likely to have hormone receptor–negative disease (24% v 7%, respectively; P < .001); DOX patients were also less likely to have pre-existing HD (29% v 45%, respectively; P < .001) or any other comorbidities at baseline (Table 1).
Chemotherapy use decreased with increasing age. Compared with patients who were 65 to 69 years old, patients who were 70 to 74, 75 to 79, and more than 80 years old were 38%, 62%, and 85% less likely to receive chemotherapy, respectively.
The propensity to receive any chemotherapy was also associated with number of comorbidities, hormone receptor negativity, mastectomy, and RT. The odds of receiving DOX increased over time, and the odds of receiving CAF decreased over time, reflecting a possible increase in the use of DOX plus cyclophosphamide. Notably, patients with pre-existing HD were less likely than other patients to receive DOX (Table 2).
Cox proportional hazards models evaluating chemotherapy and HD are shown in Table 3. Increasing age was associated with an increased rate of cardiac diseases during the study period. Compared with white patients, black patients had a 25% higher rate of HD. Patients diagnosed with HD during the year before breast cancer diagnosis were 2.4 times more likely to experience HD during follow-up. Controlling for these covariates, patients receiving chemotherapy had a 22% increase in the rate of HD, a 55% increase in the rate of CM diagnosis, and a 20% increase in the rate of CHF diagnosis compared with untreated patients. In a separate analysis of women receiving both RT and chemotherapy, the combination was associated with a slightly higher risk of HD (33%).
We also evaluated interaction terms in the models to determine what factors contributed to an increased risk of adverse cardiac outcome. We found that black patients are at highest risk of chemotherapy-induced HD (P < .007), and patients with pre-existing HD (P < .0007) and two or more comorbidities (P = .05) are at highest risk for DOX-induced CM. The interaction term in the model for RT and chemotherapy and the risk of HD was not significant (P = .84).
Patients treated with DOX were 2.5 times more likely to experience CM than untreated patients. Patients treated with CAF had higher rates of all cardiac outcomes than patients treated with CMF. The difference in hazard ratios between the two regimens was most pronounced for CM. Compared with untreated patients, CAF patients were 2.3 times more likely to experience CM, whereas CMF patients were only 1.3 times more likely to experience CM (Table 4).
The relationship between DOX use and subsequent CM was the most profound association that we found. We tested the proportionality of the hazards and found that the relationship between CM and DOX was independent of time. Chemotherapy use was not associated with an increased risk of MI. Even when using a conservative adjustment for multiple statistical tests, the P values for all of the estimates were less than the adjusted P value threshold of .007.
In the first year after diagnosis, the incidence of CM ranged from a low of 1.55% for patients not receiving chemotherapy to 4.09% for patients receiving DOX (odds ratio = 3.51; 95% CI, 2.63 to 4.69). After 5 years of follow-up, the cumulative incidence had increased to 4.97% in patients not receiving chemotherapy and 10.23% in patients receiving CAF. Figure 1A presents the annual incidence of CM by year since breast cancer diagnosis. Figure 1B presents these data cumulatively by year since diagnosis.
CM-free survival was defined as time to first CM diagnosis during the follow-up period. Figure 2 presents the Kaplan-Meier curves for CM-free survival for patients receiving CMF, DOX, and CAF compared with patients who did not receive chemotherapy. There was no difference between the Kaplan-Meier curve for patients who received CMF and the curve for patients who did not receive chemotherapy. The Kaplan-Meier curve for patients not receiving chemotherapy was significantly different from the curve for patients receiving DOX and CAF according to the log-rank test (P < .0001).
DISCUSSION
Our analysis showed that elderly women with breast cancer who had pre-existing cardiac disease were less likely than women without such disease to be treated with chemotherapy. We also found that, despite being selected for favorable cardiac risk, patients who received chemotherapy were more likely than patients who did not receive chemotherapy to experience cardiac disease subsequently. DOX use was associated with a three-fold increase in the rate of CM in the first year after treatment. As in prior studies, we found that the incidence and severity of anthracycline-induced cardiotoxicity seems to persist over time.22,51,52 Although we observed that the amplified risk of treatment-related CM decreased considerably after the first year, the risk of experiencing CM for DOX patients was still 50% higher than the risk for patients who did not receive chemotherapy in the fifth year after diagnosis.
Anthracycline-induced CM has an overall incidence of 1% to 5% that varies with the cumulative dose of anthracycline.9,48 Previous research suggests that DOX-induced cardiotoxicity may be more pronounced in the elderly. Patients more than 65 years old are 2.25 times more likely to experience CHF compared with patients less than 65 years old.49 Recently, Perez et al35 found that 23% of patients treated with four cycles of DOX plus cyclophosphamide experienced toxicity immediately after treatment, suggesting that the damage may be frequent yet reversible. Anthracycline-induced cardiotoxicity is also related to the rate of administration,50 sex,26 pre-existing HD,5 and left-sided RT.24,51,52
Our data did not include either cumulative dose of DOX or the method of administration. However, because standard treatment usually consists of doses between 240 and 360 mg/m2, it is unlikely that any patient received more than 400 mg/m2. In a prior study, we found that older patients were more likely than younger patients to terminate treatment early.53 Therefore, we consider it likely that our patients received relatively low cumulative doses. If so, our estimates of the cardiac effects of DOX should be viewed as conservative. Although our study does not provide insight into the debate on threshold dose and optimal schedule for anthracycline-induced cardiotoxicity, it does capture baseline risk factors and real-world utilization of chemotherapy in a population-based cohort of elderly breast cancer patients suffering comorbidities.
Previous studies have suggested that RT is an independent risk factor for anthracycline-induced cardiotoxicity.54-56 Many women are treated with a combination of RT and chemotherapy; yet few long-term data are available on the interactive effects of these treatments.57,58 RT did not independently predict cardiac outcomes after controlling for other covariates; however, RT does seem to potentiate the effect of chemotherapy. The results of our investigation of RT and cardiotoxicity are reported separately (Doyle et al, manuscript in preparation).
A few studies have reported no increase in the risk of cardiotoxicity after anthracycline chemotherapy. Among patients treated in a clinical trial, Ganz et al found no difference in cardiac events 5 to 8 years after treatment. The findings from this study and other clinical trials have limited generalizability to our elderly study population because the studies typically exclude patients who are older than 65 years and who are diagnosed with pre-existing cardiovascular disease.59 We found that patients who received CMF also had higher risk of cardiotoxicity than patients who did not receive chemotherapy.
Others have found that older age and comorbidities are associated with less aggressive treatment in early-stage breast cancer.22,60 Indeed, our study demonstrates that patients who were older, diagnosed at an earlier stage, hormone receptor positive, or suffering from pre-existing HD or any comorbidities were less likely to receive chemotherapy than other patients. Newcomb and Carbone17 found that only 35% of patients older than 65 years were offered chemotherapy and that such patients were twice as likely as younger patients to reject chemotherapy for fear of side effects. However, data from clinical trials suggest that older women and younger women derived similar reductions in breast cancer mortality from chemotherapy, and age alone should not be a contraindication to the use of chemotherapy in older women who are in good health.61
Selection of patients for chemotherapy was undoubtedly influenced by cardiac risk. Therefore, the chemotherapy patient cohort consists of fewer patients with cardiac disease. Both the Charlson-derived comorbidity score and pre-existing HD diagnosis were associated with subsequent cardiac events. These findings are consistent with other studies that have shown that breast cancer patients suffering from comorbidities experience worse outcomes than breast cancer patients without comorbidities.62,63
The SEER-Medicare database is an invaluable tool for studying unanticipated treatment effects and long-term outcomes in a population-based sample of patients who, for various reasons, have been underrepresented in clinical trials. Given the constraints on eligibility for such trials and other barriers to trial participation, using such databases is the only way to find out how treatments work in the real world. Our study extends the findings of clinical trial research conducted among younger women to an elderly breast cancer population.
The observational design of the study and incomplete measurement of risk factors related to HD impose some limitations on the interpretation of our findings. We used propensity score analysis to control for potential confounders on the association between treatment and outcomes. Inadequate control for health-related factors may have led to an underestimate of the cardiac risks associated with chemotherapy. However, even within the selected group of patients who received chemotherapy, patients who were treated with DOX-based chemotherapy had a higher risk of cardiotoxicity than patients who received other regimens. We doubt that this association is a result of selection bias. We also believe that even a small increase in the risk of a common and serious adverse health effect, such as cardiac disease, has important public health implications.
Our study demonstrates that the elevated risk of cardiac disease associated with adjuvant chemotherapy, which has been observed among participants in clinical trials, is also present among elderly breast cancer patients in a population-based sample. Within that sample, this risk is associated with older age, is higher among patients who receive anthracyclines, and, although most pronounced in the first year after treatment, continues over time. Given the prevalence of anthracycline regimens in the current adjuvant breast cancer setting, some patients may benefit from increased cardiac monitoring. Because the effects of chemotherapy may appear late in long-term survivors of breast cancer, there is a need for new models to assess the long-term effects of treatment.
Authors' Disclosures of Potential Conflicts of Interest
The authors indicated no potential conflicts of interest.
Acknowledgment
We acknowledge the efforts of the Applied Research Branch, Division of Cancer Prevention and Population Science, National Cancer Institute; the Office of Information Services and the Office of Strategic Planning, Health Care Financing Administration; Information Management Services, Inc; and the Surveillance, Epidemiology, and End Results (SEER) Program tumor registries in the creation of the SEER-Medicare database.
NOTES
Supported by an American Society of Clinical Oncology Career Development Award (D.L.H.), K07 Award No. CA95597 from the National Cancer Institute (D.L.H.), K05 Award No. CA89155 from the National Cancer Institute (A.I.N.), and Grant No. RSGT-01-024-04-CPHPS from the American Cancer Society (A.I.N.).
This study used the linked Surveillance, Epidemiology, and End Results-Medicare database. The interpretation and reporting of these data are the sole responsibility of the authors.
Authors' disclosures of potential conflicts of interest are found at the end of this article.
REFERENCES
Levine MN, Bramwell VH, Pritchard KI, et al: Randomized trial of intensive cyclophosphamide, epirubicin, and fluorouracil chemotherapy compared with cyclophosphamide, methotrexate, and fluorouracil in premenopausal women with node-positive breast cancer: National Cancer Institute of Canada Clinical Trials Group. J Clin Oncol 16: 2651-2658, 1998
Basser L, Abraham R, To LB, et al: Cardiac effects of high-dose epirubicin and cyclophosphamide in women with poor prognosis breast cancer. Ann Oncol 10: 53-58, 1999
Sparano JA: Use of dexrazoxane and other strategies to prevent cardiomyopathy associated with doxorubicin-taxane combinations. Semin Oncol 25: 66-71, 1998
Singal PK, Iliskovic N: Doxorubicin-induced cardiomyopathy. N Engl J Med 339: 900-905, 1998
von Hoff DD, Layard MW, Basa P, et al: Risk factors for doxorubicin-induced congestive heart failure. Ann Intern Med 91: 710-717, 1979
Zhou S, Starkov A, Froberg MK, et al: Cumulative and irreversible cardiac mitochondrial dysfunction induced by doxorubicin. Cancer Res 61: 771-777, 2001
Lefrak EA, Pitha J, Rosenheim S, et al: A clinicopathologic analysis of Adriamycin cardiotoxicity. Cancer 32: 302-314, 1973
Gewirtz DA: A critical evaluation of the mechanisms of action proposed for the antitumor effects of the anthracycline antibiotics adriamycin and daunorubicin. Biochem Pharmacol 57: 727-741, 1999
Hortobagyi GN: Anthracyclines in the treatment of cancer: An overview. Drugs 54: 1-7, 1997
Zambetti M, Moliterni A, Materazzo C, et al: Long-term cardiac sequelae in operable breast cancer patients given adjuvant chemotherapy with or without doxorubicin and breast irradiation. J Clin Oncol 19: 37-43, 2001
Crivellari D, Bonetti M, Castiglione-Gertsch M, et al: Burdens and benefits of adjuvant cyclophosphamide, methotrexate, and fluorouracil and tamoxifen for elderly patients with breast cancer: The International Breast Cancer Study Group Trial VII. J Clin Oncol 18: 1412-1422, 2000
Fumoleau P, Kerbrat P, Romestaing P, et al: Randomized trial comparing six versus three cycles of epirubicin-based adjuvant chemotherapy in premenopausal, node-positive breast cancer patients: 10-Year follow-up results of the French Adjuvant Study Group 01 Trial. J Clin Oncol 21: 298-305, 2003
Hutchins LF, Unger JM, Crowley JJ, et al: Underrepresentation of patients 65 years of age or older in cancer-treatment trials. N Engl J Med 341: 2061-2067, 1999
Extermann M, Balducci L, Lyman GH: What threshold for adjuvant therapy in older breast cancer patients? J Clin Oncol 18: 1709-1717, 2000
Hillner BE, Penberthy L, Desch CE, et al: Variation in staging and treatment of local and regional breast cancer in the elderly. Breast Cancer Res Treat 40: 75-86, 1996
Mandelblatt JS, Hadley J, Kerner JF, et al: Patterns of breast carcinoma treatment in older women: Patient preference and clinical and physical influences. Cancer 89: 561-573, 2000
Newcomb PA, Carbone PP: Cancer treatment and age: Patient perspectives. J Natl Cancer Inst 85: 1580-1584, 1993
Chu J, Diehr P, Feigl P, et al: The effect of age on the care of women with breast cancer in community hospitals. J Gerontol 42: 185-190, 1987
Osteen RT, Cady B, Chmiel JS, et al: 1991 national survey of carcinoma of the breast by the Commission on Cancer. J Am Coll Surg 178: 213-219, 1994
Harlan LC, Abrams J, Warren JL, et al: Adjuvant therapy for breast cancer: Practice patterns of community physicians. J Clin Oncol 20: 1809-1817, 2002
Du XL, Key CR, Osborne C, et al: Discrepancy between consensus recommendations and actual community use of adjuvant chemotherapy in women with breast cancer. Ann Intern Med 138: 90-97, 2003
Du X, Goodwin JS: Patterns of use of chemotherapy for breast cancer in older women: Findings from Medicare claims data. J Clin Oncol 19: 1455-1461, 2001
Hebert-Croteau N, Brisson J, Latreille J, et al: Time trends in systemic adjuvant treatment for node-negative breast cancer. J Clin Oncol 17: 1458-1464, 1999
Bristow MR, Billingham ME, Mason JW, et al: Clinical spectrum of anthracycline antibiotic cardiotoxicity. Cancer Treat Rep 62: 873-879, 1978
Hrushesky WJ, Fader DJ, Berestka JS, et al: Diminishment of respiratory sinus arrhythmia foreshadows doxorubicin-induced cardiomyopathy. Circulation 84: 697-707, 1991
Lipshultz SE, Lipsitz SR, Mone SM, et al: Female sex and drug dose as risk factors for late cardiotoxic effects of doxorubicin therapy for childhood cancer. N Engl J Med 332: 1738-1743, 1995
Silber JH, Barber G: Doxorubicin-induced cardiotoxicity. N Engl J Med 333: 1359-1360, 1995
Krischer J, Epstein S, Cuthbertson D, et al: Clinical cardiotoxicity following anthracycline treatment for childhood cancer: The Pediatric Oncology Group experience. J Clin Oncol 15: 1544-1552, 1997
Bristow MR, Thompson PD, Martin RP, et al: Early anthracycline cardiotoxicity. Am J Med 65: 823-832, 1978
Shan K, Lincoff AM, Young JB: Anthracycline-induced cardiotoxicity. Ann Intern Med 125: 47-58, 1996
Grenier MA, Lipshultz SE: Epidemiology of anthracycline cardiotoxicity in children and adults. Semin Oncol 25: 72-85, 1998
Steinberg JS, Cohen AJ, Wasserman AG, et al: Acute arrhythmogenicity of doxorubicin administration. Cancer 60: 1213-1218, 1987
von Hoff DD, Rozencweig M, Piccart M: The cardiotoxicity of anticancer agents. Semin Oncol 9: 23-33, 1982
Allen A: The cardiotoxicity of chemotherapeutic drugs. Semin Oncol 19: 529-542, 1992
Perez EA, Suman VJ, Davidson NE, et al: Effect of doxorubicin plus cyclophosphamide on left ventricular ejection fraction in patients with breast cancer in the North Central Cancer Treatment Group N9831 Intergroup Adjuvant Trial. J Clin Oncol 22: 3700-3704, 2004
Freter CE, Lee TC, Billingham ME, et al: Doxorubicin cardiac toxicity manifesting seven years after treatment: Case report and review. Am J Med 80: 483-485, 1986
Steinherz LJ, Steinherz PG, Tan CT, et al: Cardiac toxicity 4 to 20 years after completing anthracycline therapy. JAMA 266: 1672-1677, 1991
Gottlieb SL, Edmiston WA Jr, Haywood LJ: Late, late doxorubicin cardiotoxicity. Chest 78: 880-882, 1980
Warren JL, Klabunde CN, Schrag D, et al: Overview of the SEER-Medicare data: Content, research applications, and generalizability to the United States elderly population. Med Care 40: IV–3-IV–18, 2002 (suppl 8)
Warren JL, Harlan LC, Fahey A, et al: Utility of the SEER-Medicare data to identify chemotherapy use. Med Care 40: IV–55-IV–61, 2002 (suppl 8)
Du X, Freeman JL, Goodwin JS: Information on radiation treatment in patients with breast cancer: The advantages of the linked medicare and SEER data—Surveillance, Epidemiology and End Results. J Clin Epidemiol 52: 463-470, 1999
Klabunde CN, Potosky AL, Legler JM, et al: Development of a comorbidity index using physician claims data. J Clin Epidemiol 53: 1258-1267, 2000
Charlson ME, Pompei P, Ales KL, et al: A new method of classifying prognostic comorbidity in longitudinal studies: Development and validation. J Chronic Dis 40: 373-383, 1986
Extermann M, Overcash J, Lyman GH, et al: Comorbidity and functional status are independent in older cancer patients. J Clin Oncol 16: 1582-1587, 1998
Du XL, Osborne C, Goodwin JS: Population-based assessment of hospitalizations for toxicity from chemotherapy in older women with breast cancer. J Clin Oncol 20: 4636-4642, 2002
D'Agostino RB: Tutorial in biostatistics: Propensity score methods for bias reduction in the comparison of a treatment to a non-randomized control group. Stat Med 17: 2265-2281, 1998
Rubin DB: Estimating causal effects from large data sets using propensity scores. Ann Intern Med 127: 757-763, 1997
Hahm S, Dresner HS, Podwall D, et al: DNA biomarkers antecede semiquantitative anthracycline cardiomyopathy. Cancer Invest 21: 53-67, 2003
Swain SM, Whaley FS, Ewer MS: Congestive heart failure in patients treated with doxorubicin: A retrospective analysis of three trials. Cancer 97: 2869-2879, 2003
Legha SS, Benjamin RS, Mackay B, et al: Reduction of doxorubicin cardiotoxicity by prolonged continuous intravenous infusion. Ann Intern Med 96: 133-139, 1982
Praga C, Beretta G, Vigo PL, et al: Adriamycin cardiotoxicity: A survey of 1273 patients. Cancer Treat Rep 63: 827-834, 1979
Pihkala J, Saarinen UM, Lundstrom U, et al: Myocardial function in children and adolescents after therapy with anthracyclines and chest irradiation. Eur J Cancer 32A: 97-103, 1996
Hershman D, McBride R, Jacobson J, et al: Racial disparities in treatment and survival among women with early-stage breast cancer. J Clin Oncol 23: 6639-6646, 2005
Eltringham RJ, Fajardo L, Stewart RJ, et al: Investigation of cardiotoxicity in rabbits from Adriamycin and fractionated cardiac irradiation preliminary result. Front Radiat Ther Oncol 13: 21-35, 1979
Ryberg M, Nielsen D, Skovsgaard T, et al: Epirubicin cardiotoxicity: An analysis of 469 patients with metastatic breast cancer. J Clin Oncol 16: 3502-3508, 1998
Hardenbergh PH, Recht A, Gollamudi S, et al: Treatment-related toxicity from a randomized trial of the sequencing of doxorubicin and radiation therapy in patients treated for early stage breast cancer. Int J Radiat Oncol Biol Phys 45: 69-72, 1999
Shapiro CL, Recht A: Side effects of adjuvant treatment of breast cancer. N Engl J Med 344: 1997-2008, 2001
Recht A, Come SE, Henderson IC, et al: The sequencing of chemotherapy and radiation therapy after conservative surgery for early-stage breast cancer. N Engl J Med 334: 1356-1361, 1996
Ganz PA, Green SJ, Hutchins L, et al: Late cardiac effects of adjuvant CMF vs CAF in women with node positive breast cancer treated on SWOG 8897: Initial results from SWOG 9342. Presented at the San Antonio Breast Cancer Symposium, San Antonio, TX, December 11-14, 2002
Newschaffer CJ, Penberthy L, Desch CE, et al: The effect of age and comorbidity in the treatment of elderly women with nonmetastatic breast cancer. Arch Intern Med 156: 85-90, 1996
Muss HB, Woolf SH, Cirrincione C, et al: Adjuvant chemotherapy in older and younger women with lymph node-positive breast cancer. JAMA 293: 1073-1081, 2005
Yancik R, Wesley MN, Ries LA, et al: Effect of age and comorbidity in postmenopausal breast cancer patients aged 55 years and older. JAMA 285: 885-892, 2001
Read WL, Tierney RM, Page NC, et al: Differential prognostic impact of comorbidity. J Clin Oncol 22: 3099-3103, 2004
Submitted May 2, 2005; accepted September 13, 2005.(John J. Doyle, Alfred I. )