Getting to the Heart of the Matter

Early successes in the treatment of childhood cancer mean that some individuals are now surviving 40 to 50 years after diagnosis, and late adverse effects of therapy can be quantified. Cardiac dysfunction is one of the most troubling cancer treatment–related late effects and is most often associated with the use of anthracycline chemotherapy and cardiac irradiation. In an article that accompanies this editorial, a group of Dutch investigators describes the incidence of cardiac events in a population-based series of 5-year survivors of childhood cancer (N 1,362) who were treated at a single institution between 1966 and 1996. The authors report 50 cardiac events, including 27 cases of cardiac failure in 42 survivors that occurred at a median attained age of 27.1 years. Similar to other studies, this article demonstrates an increasing risk of experiencing an adverse cardiac event with increasing exposure to anthracyclines and cardiac irradiation. Study authors have verified the cardiac events, and follow-up of survivors is almost complete; these are important factors in reliably estimating the frequency of events. A previous investigation from the Childhood Cancer Survivor Study reported a somewhat higher incidence of cardiac events than that observed in the current study (4.1% at 30 years in the Childhood Cancer Survivor Study compared with 2.7% in the Dutch study) in a large cohort of 5-year-plus survivors of cancer. This difference was perhaps a result of a selection bias that is commonly seen in nonpopulation-based studies in which survivors with problems are more likely to participate. The potential adverse impact of cardiac toxicity on long-term survival is supported by the Dutch report of poor outcomes for the survivors with cardiac events; only four survivors recovered fully. Among children who were treated with both anthracyclines and cardiac irradiation, one in eight had severe heart disease 30 years later. Although the frequency of severe cardiac complications is alarming, all studies that examine truly late outcomes (beyond 20 years) suffer from the problem that they observe and report outcomes of therapies that have since been modified and in some instances are no longer used. For example, during the last 20 years, therapy has been significantly reduced in children with Hodgkin’s lymphoma to diminish late adverse effects of a highly curable malignancy. However, these new data are useful for quantifying risk and recommending follow-up for the cohort of survivors now in young adulthood and entering middle age, most of whom do not have access to late-effects specialists in multidisciplinary survivorship clinics. Identification of individuals at highest risk of treatment-related morbidity is valuable so that aggressive follow-up efforts can be targeted at those with the most to gain. This information can also be used to support the need for health care resources and insurance coverage of appropriate health screening of young adult survivors and guide the development of safer treatment approaches for newly diagnosed patients. In regard to the latter, two additional articles accompany this editorial and address individual susceptibility to cardiac damage after anthracycline exposure. Marked heterogeneity in individual susceptibility is evident, given that some children develop cardiac toxicity after exposure to anthracycline doses as low as 100 mg/m, whereas others can receive more than 500 mg/m without problems. Visscher et al performed a broad candidate gene study of 2,977 singlenucleotide polymorphisms (SNPs) in 220 key drug biotransformation genes in a discovery cohort of 156 children from Canada who were treated with anthracycline. The finding of a number (nine) of SNPs that were significantly associated with cardiac toxicity in this initial cohort was replicated in a second cohort of 188 children from across Canada, and additionally in a third cohort of 96 children from the Netherlands. Successful replication of data in this study is a major strength, and was achieved in cohorts that are small for a study of this kind. The use of a candidate gene approach, rather than a genomewide strategy, adds biologic plausibility to the findings. Many prior pharmacogenetic studies have been hampered by a lack of reproducibility, perhaps because observations of a small effect size in a setting of multiple comparisons can occur by chance or can be specific for the dosage and treatment regimen under study. A second pharmacogenetic study from the North American Children’s Oncology Group (COG) used a case (survivors with cardiomyopathy; n 170) control (survivors matched on length of follow-up, cancer diagnosis, year of diagnosis, and race, without cardiomyopathy; n 317) design. The authors determined that a polymorphism in the gene CBR3 predicted risk of cardiomyopathy in children exposed to low to moderate doses of anthracyclines ( 250 mg/m), although exposure to high doses of anthracyclines ( 250 mg/m) was associated with increased risk irrespective of CBR3 genotype. Disconcertingly, this same polymorphism was included in the study by Visscher et al and was not associated with risk of cardiac events. The reason for this discrepancy is not immediately evident, although the end points of the two studies were modestly different. Visscher et al included all cardiac events, including left ventricular dysfunction measured by echocardiography (shortening fraction 26%) or symptoms requiring intervention, and controls all had a normal echocardiogram more than 5 years after completion of therapy. Cardiomyopathy was defined in the COG study as symptoms or signs of cardiac decompensation, or echocardiographic evidence of left ventricular dysfunction as evidenced by an ejection fraction 40% or shortening fraction 28%, a relatively minor difference JOURNAL OF CLINICAL ONCOLOGY E D I T O R I A L VOLUME 30 NUMBER 13 MAY 1 2012