The Heart of the Matter

Abstract

Although anthracyclines have served as the mainstay of effective cytotoxic therapy for breast cancer during the last 30 years, the time has come to evaluate whether the benefits outweigh the significant risks of cardiac toxicity. In this issue of the Journal of Clinical Oncology, Pinder et al report an increased incidence of congestive heart failure (CHF) in older women who had more than 10 years follow-up and were treated with adjuvant anthracyclines for breast cancer. Women who received anthracyclines had higher rates of CHF, even though they were younger and had fewer comorbidities compared with women who received nonanthracycline regimens or no chemotherapy. For women aged 66 to 70 years at 10 years follow-up, a diagnosis of CHF was made in 38.4%, 32.5%, and 29% for anthracyclinetreated, nonanthracycline-treated, and no-chemotherapy groups, respectively. Age, African American race, hypertension, diabetes, and coronary artery disease were significant predictors of CHF. Although follow-up was shorter and with fewer patients, trastuzumab treatment was a significant predictor of CHF. These results further emphasize the magnitude of longterm risk for cardiovascular disease in older patients treated with anthracyclines, which has been previously reported by others. The observation of ethnicity as an independent risk factor for CHF highlights the importance of population studies, because these patients are under-represented in clinical trials. In this study, African American women had a relative 49% higher risk of developing CHF compared with white women, but they also had more advanced disease at diagnosis and were more likely to receive anthracycline-containing chemotherapy. Identification of at-risk populations will help to tailor therapies aimed specifically to reduce anthracycline-related cardiac toxicities. However, like all population-based studies, this study has inherent limitations and weaknesses. The database used in this study was compiled for reimbursement purposes on the basis of nonrandomly selected populations. The Surveillance, Epidemiology, and End Results (SEER)-Medicare captures only services covered by Medicare and does not include information on health maintenance organization enrollees, who account for a significant percentage of the Medicare population. In addition, records of certain conditions are inaccurate because of inadequate coding. Therefore, when one interprets population studies, limitations of the data and potential bias of data sampling should be taken into consideration. Some of the specific concerns are as follows, and most of them are pointed out by the authors. First, because no information was available to determine the severity and diagnostic criteria for CHF or to discriminate between systolic and diastolic dysfunction, the potential for misdiagnosis of patients with mild or nonclassic symptoms was substantial. Second, critical information, such as cumulative anthracycline dose, other toxicities, and hormonal treatment, was not captured in the database. It should be noted that, on the basis of a study in which left ventricular ejection fraction was obtained carefully and frequently, cardiotoxicity occurred at doxorubicin cumulative doses of 300 mg/m and lower. This is much lower than a previously recommended threshold dose of 500 mg/m. Third, the link between radiation and CHF may have been underestimated because of the lack of specific information on the techniques used and the dose of radiation to the heart. A recent report found that radiation to the internal mammary chain or a higher mean dose of radiation to the heart was associated with increased risk of CHF. This association was not apparent if radiation alone administered to the left or right side was independently considered. These limitations may have contributed to the surprising observation that anthracyclines did not significantly increase the risk of CHF in women ages 71 to 80 years. This finding directly contradicts the results from other studies suggesting that anthracycline-associated cardiac toxicity increases with increased age. In addition, by excluding other cardiac diseases and focusing on only CHF, this study may have missed an even broader scope of cardiovascular toxicity induced by anthracyclines, especially in this older population. Other factors that may have potential for causing long-term cardiac toxicity should also be considered. The rates of CHF reported in this study were high compared with CHF prevalence in the general population. It may have been useful to have a control group without breast cancer from the Medicare database. Increased long-term risk for CHF has been reported in patients who have received cyclophosphamide, methotrexate, and fluorouracil (CMF). Also, preliminary information from one trial with an aromatase inhibitor, letrozole, reported an increase in ischemia and CHF in the letrozole-treated group, although this was not statistically significant. Evidence of cardiac toxicity from other treatment modalities of breast cancer may emerge from longer follow-up. Despite its limitations, this study provides valuable information to the current debate on the role of adjuvant anthracyclines. JOURNAL OF CLINICAL ONCOLOGY E D I T O R I A L VOLUME 25 NUMBER 25 SEPTEMBER 1 2007