In the United States, where 1 in 680 people between 20 and 50 years old are survivors of childhood cancer, the impact of long-term health consequences is a cause for concern, and even more so because this population is increasing.
Treatment advances have extended the lives of children with cancer. Between 2004 and 2010, the survival rate in the United States in children up to 14 years of age with all types of cancer was 83%— significantly higher than the 58% reported in the mid-1970s.1 Furthermore, between 1970 and 2011, the death rates from cancers diagnosed in children up to 14 years old, and in adolescents 15 to 19 years old, declined by 67% (from 6.3 to 2.1 per 100,000 people) and 58% (from 7.2 to 3.0 per 100,000 people), respectively.1 However, being cured of cancer is usually not without consequences. Within the first 30 years after diagnosis, survivors of childhood cancers have approximately a 75% cumulative incidence of treatment-related chronic health problems.2 In the United States, where 1 in 680 people between 20 and 50 years old are survivors of childhood cancer,3 the impact of long-term health consequences is a cause for concern, and even more so because this population is increasing.
Some commonly used cancer drugs, such as the anthracyclines, are known to be cardiotoxic. Left undetected and untreated, this cardiotoxicity is progressive and persistent and can lead to cardiomyopathy, clinical heart failure, the need for a heart transplant, or death.4 In fact, 30 years after diagnosis, the number of cardiac-related deaths among survivors exceeds the number caused by cancer recurrence.5 The prevalence of cardiovascular events, even 5 years after diagnosis, is higher in survivors than in healthy controls (see Figure 1). In addition, compared with controls, survivors are:
· Fifteen times as likely to have heart failure2
· Ten times as likely to have coronary artery disease2
· Nine times as likely to have a cerebrovascular event2
· Eight times as likely to die from cardiovascular-related disease.6
Chemotherapeutic agents may cause adverse cardiac effects either directly, by compromising myocardial structure and function, or indirectly, by impairing vascular hemodynamics or other organ systems such as the endocrine glands, which may result in endocrinopathies. However, pediatric drug toxicity cannot be predicted based on observation of adult patients.7 While several cardiovascular toxicity studies have been conducted in adult cancer patients, far fewer have been conducted in pediatric cancer patients. Hence, many pediatric treatment protocols are extrapolated from those for adults, which is not always appropriate given the differences in body composition and developmental changes in children. For example, early cardiotoxicity in adults may be lower when anthracyclines are administered as a continuous infusion than as a bolus infusion. However, evidence in children with high-risk acute lymphoblastic leukemia (ALL) indicates that a continuous infusion is not more cardioprotective than a bolus infusion.8 These results suggest that a continuous infusion in children does not afford incremental oncologic efficacy, but entails the added expense of longer hospital stays and the increased risk of complications, suggesting that continuous infusion of anthracyclines in children for cardioprotection should be contraindicated until evidence to the contrary emerges.
Multiple risk factors for cardiovascular toxicity during and after treatment have been identified. These factors include the cumulative dose of anthracycline, concomitant radiation therapy, younger age at diagnosis, female sex, black race, and the presence of other cardiovascular comorbidities.9,10 Despite these risk factors, the occurrence of cardiotoxicity remains variable in children, indicative of genetic predisposition.
Monitoring the cardiovascular status of children treated with chemotherapy might detect early cardiotoxicity, even when left ventricular (LV) dysfunction is asymptomatic, thus providing opportunities to prevent, reduce, or treat the condition before it worsens.
Echocardiography is commonly used to monitor cardiac structure and function in anthracycline-treated, long-term survivors of childhood cancer. It is noninvasive, painless, and widely available, and therefore convenient. The Children’s Oncology Group has published recommended guidelines for long-term cardiovascular monitoring.11 Following these guidelines and acting on subsequent abnormal findings could theoretically result in an incremental cost-effectiveness ratio of $61,500, extend life expectancy by 6 months, and improve quality-adjusted life-years by 1.6 months. Additionally, it could, in theory, reduce the cumulative incidence of heart failure by 18% at 30 years after cancer diagnosis.12
In contrast, a simulation in which patients were categorized either as low-risk for anthracycline cardiotoxicity (defined by a cumulative anthracycline dose <250 mg/m2) or high-risk (a cumulative anthracycline dose ≥250 mg/m2), but were not followed with specific cardiovascular monitoring guidelines,13 found an overall 18.8% lifetime risk for systolic heart failure in 5-year survivors of childhood cancer aged 15 years, with an average age at onset of 58 years. Further, cardiac assessments and subsequent monitoring-directed treatment every 10 years theoretically reduced the lifetime risk by 2.3%, while a yearly assessment and subsequent treatment reduced the risk by 8.7%—the model predicted incremental cost-effectiveness ratios of $111,600 and $278,600, respectively.13 These 2 studies illustrate the challenges with establishing reliable theoretical evidence for such guidelines. Furthermore, the models were restricted to the monitoring of long-term survivors; however, the ability to implement cardiovascular guidelines to predict long-term cardiac outcomes, and of biomarker-guided dose modification to improve the overall outcome—defined as the quality of life for a child with cancer and their family over their lifespan, to maximize treatment efficacy and minimize toxicity and late effect outcomes—is yet unknown.
Echocardiography, however, lacks the sensitivity and specificity to detect early subclinical abnormalities of LV structure and function in survivors of childhood cancer. Both LV ejection fraction and LV fractional shortening are load-dependent, cannot reliably detect restrictive anthracycline-related cardiomyopathy, and may not identify changes in load-independent LV contractility. Thus, abnormalities in these measurements recorded during therapy may result from causes unrelated to anthracycline-induced myocardial injury.14
Newer imaging techniques are being explored but have not been fully adopted by pediatric oncologists, given limited evidence of sensitivity, specificity, and safety in children. For example, Doppler speckle-tracking—derived longitudinal strain echocardiography has been useful in assessing cardiac damage in adults with,15 and without, cancer,16 but it has not been studied in children treated with chemotherapy to the point were it could be validated as a surrogate outcome for late cardiotoxicity in long-term survivors.17,18 Cardiac magnetic resonance imaging may provide quality images of LV function in echo-poor windows, such as in obese patients, but it is expensive, time-consuming, not widely available, requires a trained physician to interpret the results, and may require sedating younger patients.18 In children treated with chemotherapy, there is still no validated method of imaging during therapy for predicting late, clinically important cardiovascular disease. Furthermore, the impact of these newer techniques in routine surveillance, and the optimal timing and cost-effectiveness for such monitoring, requires further investigation.18,19
Interest in the use of the serum biomarkers such as cardiac troponin-T (cTnT), cardiac troponin-I (cTnI), and N-terminal pro-brain natriuretic peptide (NT-proBNP), as an additional means of evaluating cardiotoxicity during and after chemotherapy, is growing. Elevated concentrations of cTnT and cTnI, which are intra-cardiomyocyte contractile proteins detectable in blood after active cardiomyocyte injury or necrosis, generally indicate irreversible cardiomyocyte loss.20,21 Concentrations of NT-proBNP, a nonspecific marker of ventricular wall stress, can be elevated in several cardiovascular conditions, including cardiomyopathy with increased LV wall stress from pressure or volume overload and heart failure. Increased concentrations of these biomarkers are associated with late adverse cardiac outcomes, as identified by echocardiography, in children receiving anthracyclines for high-risk ALL.22 For example, elevated cTnT concentrations during the first 90 days of doxorubicin therapy were associated with reduced LV mass and LV end-diastolic posterior wall thickness-to-dimension ratio, a marker of pathologic LV remodeling, 4 years later.22 Similarly, elevated NT-proBNP concentrations during the first 90 days of therapy were associated with an abnormal LV thickness-dimension ratio, suggesting pathologic LV remodeling, 4 years later.22 Other cardiac biomarkers indicative of the development or progression of heart failure have not been validated as surrogates of late cardiac status in long-term survivors of childhood cancer treated with anthracyclines.
Drawing on known or potential risk factors, investigators have studied several methods to reduce the cardiac complications of anthracyclines.23 Because the most prominent risk factor is the cumulative dose of anthracycline, protocols over the past several decades have tested the efficacy of lower cumulative doses. In the 1970s, before the cardiotoxicity of anthracycline was known, childhood ALL clinical trials would administer cumulative doses of doxorubicin greater than 400 mg/m2. Several years later, these patients experienced persistent and progressive, clinically important adverse LV effects.9,24 As a result, subsequent protocols in the 1980s and early 1990s reduced the cumulative doses of anthracycline to 45 to 60 mg/m2 for children with standard-risk ALL, and to 345 to 360 mg/m2 for children with high-risk ALL.25
Despite a lower risk of cardiotoxicity with the reduced cumulative doses of anthracycline, children with high-risk ALL remained at increased risk for late LV abnormalities.8,26 In the 1990s, an analysis of 189 long-term survivors of ALL from the Dana-Farber Cancer Institute ALL Consortium and patients treated in Denmark revealed a lower risk of LV abnormalities in survivors who received a cumulative doxorubicin dose of ≤300 mg/m2 than in those who received >300 mg/m2 after a median follow-up of 8 years.27 Thus, the cumulative dose for high-risk ALL protocols from 1995 onward was again reduced to 300 mg/m2.25 Although reducing cumulative anthracycline doses may offer some cardioprotection, it may also reduce treatment efficacy.26 In addition, although doses of ≤300 mg/m2 reduce the risk of cardiotoxicity, they do not eliminate the risk.28 Subclinical cardiac abnormalities have been detected at even the smallest doses of anthracyclines (≤100 mg/m2), almost 10 years after diagnosis.29 In reality, there is no safe dose of anthracyclines for children with cancer if the goal is to avoid lifetime cardiac abnormalities.33
The most promising solution for preventing cardiotoxicity is the coadministration of dexrazoxane. Dexrazoxane is a chelating agent that reduces the formation of anthracycline-iron complexes, thus interfering with iron-mediated free radical generation.30,31 It also mitigates doxorubicin-induced DNA damage by inhibiting topoisomerase 2-beta.32 Dexrazoxane is currently approved by the FDA for use in adults with metastatic breast cancer who have received a cumulative dose of 300 mg/m2 of doxorubicin and who may benefit from continued treatment with an anthracycline. In August 2014, doxorubicin was designated by the FDA as a drug for orphan diseases in children.33 Studies in children with high-risk cancer have documented the cardioprotective effects of dexrazoxane when administered before each dose of doxorubicin.33
Among 206 children with high-risk ALL randomly assigned to receive dexrazoxane before each dose of doxorubicin (dexrazoxane group) or doxorubicin alone, the number of children with elevations in serum cTnT concentrations—from diagnosis to the end of doxorubicin treatment—was lower in the dexrazoxane group than in the doxorubicin-only group (21% vs 50%) (see Figure 2). However, about 2 months after doxorubicin treatment, LV fractional shortening and contractility were depressed in both treatment groups, which suggests that echocardiographic measurements are not valid surrogates for subclinical myocardial injury in this setting.34 A follow-up study of this same cohort, 5 years after completing doxorubicin treatment, showed significantly abnormal mean z scores for LV fractional shortening and end-systolic dimension in the doxorubicin-alone group, but not the dexrazoxane group.28 Left ventricular wall thickness and thickness-to-dimension ratio differed significantly between groups of treated girls. Furthermore, in a planned subgroup analysis after 5 years, girls receiving dexrazoxane had better values than boys for LV end-diastolic thickness-to-dimension ratio (a marker of pathologic LV remodeling) and LV fractional shortening.28 Similar findings of cardioprotection with dexrazoxane have been reported in other large studies of children treated with doxorubicin for T-cell ALL and lymphoma35; in children with osteosarcoma,36 (C.L. Schwartz, MD, MPH; L.H. Wexler, MD; M. Devidas, PhD, et al. Unpublished results.) whose treatment also included the known cardiotoxic drug, trastuzumab; and in children with osteosarcoma treated with doxorubicin dose escalations up to 600 mg/m2 cumulative dose. (C.L. Schwartz, MD, MPH; L.H. Wexler, MD; M. Devidas, PhD, et al. Unpublished results.)
Despite favorable evidence that dexrazoxane offers cardioprotection in children who receive anthracyclines as part of their treatment, only 2% of children with acute myeloid leukemia (AML) and ALL in the United States received dexrazoxane in clinical practice between 1999 and 2009.37 Physicians have been reluctant to use dexrazoxane for fear of increasing the risk of second malignancies or recurrence. However, several large observational studies have found no evidence of such risk. Significant differences have not been found in 5-year event-free survival (77% for the doxorubicin-only group vs 76% for the dexrazoxane group) or in the incidence of secondary malignancies or recurrence.28,38-40
Advances in cancer treatments have brought hope to children with a diagnosis of cancer—a disease once deemed incurable. However, years later, these same treatments can result in a lowering of quality of life, in part due to cardiotoxicity associated with the treatments. Because treatment-induced cardiotoxicity can be pervasive, persistent, and progressive, developing evidence-based surveillance guidelines is important to identify early subclinical abnormalities that, together with validated risk profiles, can help guide treatment to maximize oncologic efficacy and minimize the risk of long-term morbidity.
Substantial evidence indicates that dexrazoxane is a safe and effective cardioprotective drug in children treated with anthracyclines. As investigators continue to search for less-cardiotoxic drugs and more effective prevention and treatment strategies, we encourage the use of dexrazoxane in protocols involving anthracycline-based chemotherapy in treating childhood cancers such as AML and high-risk ALL. In fact, dexrazoxane has become part of the standard of care for patients in the Children’s Oncology Group and the Dana-Farber Cancer Institute Childhood ALL clinical trials.33 Most essential of all is that cardiologists and oncologists collaborate to find treatments that balance oncologic efficacy with the risks of cardiotoxicity to maximize the quality of life and survival for long-term survivors of childhood cancer.
Vivian I. Franco, MPH, is research coordinator at Wayne State University School of Medicine.
Steven E. Lipshultz, MD, is pediatrician-in-chief, Children's Hospital of Michigan, and chair of pediatrics, Wayne State University School of Medicine.