Impact of hemochromatosis gene mutations on cardiac status in doxorubicin-treated survivors of childhood high-risk leukemia


  • We thank all of the patients and families involved in this study, as well as the data managers and study personnel from the 8 participating centers for this study in the Dana-Farber Cancer Institute Consortium.



Doxorubicin is associated with progressive cardiac dysfunction, possibly through the formation of doxorubicin-iron complexes leading to free-radical injury. The authors determined the frequency of hemochromatosis (HFE) gene mutations associated with hereditary hemochromatosis and their relationship with doxorubicin-associated cardiotoxicity in survivors of childhood high-risk acute lymphoblastic leukemia.


Peripheral blood was tested for 2 common HFE allelic variants: C282Y and H63D. Serum cardiac troponin-T (cTnT) and N-terminal pro-brain natriuretic peptide (NT-proBNP), which are biomarkers of cardiac injury and cardiomyopathy, respectively, were assayed during therapy. Left ventricular (LV) structure and function were assessed with echocardiography.


A total of 184 patients had DNA results for at least 1 variant, and 167 had DNA results for both: 24% carried H63D and 10% carried C282Y. Heterozygous C282Y genotype was associated with multiple elevations in cTnT concentrations (P = .039), but not NT-proBNP. At a median of 2.2 years (range, 1.0 years–3.6 years) after diagnosis, the mean Z-scores for LV fractional shortening (−0.71 [standard error (SE), 0.25]; P = .008), mass (−0.84 [SE, 0.17]; P < .001), and end-systolic (−4.36 [SE, 0.26], P < .001) and end-diastolic (−0.68 [SE, 0.25]; P = .01) posterior wall thickness were found to be abnormal in children with either allele (n = 32). Noncarriers (n = 63) also were found to have below-normal LV mass (−0.45 [SE, 0.15]; P = .006) and end-systolic posterior wall thickness (−4.06 [SE, 0.17]; P < .001). Later follow-up demonstrated similar results.


Doxorubicin-associated myocardial injury was associated with C282Y HFE carriers. Although LV mass and wall thickness were found to be abnormally low overall, they were even lower in HFE carriers, who also had reduced LV function. Screening newly diagnosed cancer patients for HFE mutations may identify those at risk for doxorubicin-induced cardiotoxicity. Cancer 2013;119:3555–3562.. © 2013 American Cancer Society.


Children with acute lymphoblastic leukemia (ALL) are reported to have long-term event-free survival rates of > 80%.[1] Late effects often include anthracycline-associated cardiovascular abnormalities.

Among more than 300,000 childhood cancer survivors in the United States,[2] > 50% have been treated with anthracyclines.[3] Anthracyclines have been associated with progressive cardiotoxicity.[4] Thus, understanding the causes, mechanisms, and magnitude of doxorubicin-related cardiotoxicity is important. Reducing cardiotoxicity could improve long-term survivor outcomes.

The risk of doxorubicin-associated cardiotoxicity is greater in girls, those with Down syndrome, and those treated at younger ages, with higher doxorubicin dose rates and cumulative doses, as well as with longer follow-up since receiving doxorubicin.[4-6] However, some patients appear to be more vulnerable than others, independent of these risk factors. Identifying those patients at the highest risk of developing late cardiotoxicity is a priority because they should be the focus of alternative chemotherapies[7] and novel preventive treatments to maintain oncologic efficacy while reducing toxicity and late effects.

Doxorubicin-induced cardiotoxicity is caused, in part, by myocardial doxorubicin-iron complexes generating doxorubicin semiquinone free radicals, which after reacting with oxygen lead to lipid peroxidation and DNA damage.[8, 9] The cardioprotective properties of dexrazoxane, a compound that chelates iron thereby reducing this free radical formation, when administered before doxorubicin supports this hypothesis.[10, 11]

We hypothesized that conditions leading to higher iron tissue concentrations may favor the development of doxorubicin cardiotoxicity. In vitro and animal studies have suggested that tissue iron-loading potentiates anthracycline cardiotoxicity, and increases biomarker concentrations of cellular damage.[8, 12]

Hereditary hemochromatosis (HH) is a genetic iron metabolism disorder resulting in iron overload-associated tissue injury. The most common HH disease-associated gene relates to specific alleles of the hemochromatosis gene (HFE); the majority of affected individuals carry at least 1 copy of a founder mutation, C282Y. Homozygosity for the C282Y mutation is present in approximately 52% to 100% of patients with HH.[13] Adults with non–cancer-associated idiopathic dilated cardiomyopathy and early pathologic left ventricular (LV) remodeling[14] are reported to have higher C282Y homozygosity than healthy controls.[15]

Postmortem testing of an anthracycline-naïve patient aged 14 years with high-risk ALL, iron overload, hepatic failure, and cardiac dysfunction revealed a homozygous C282Y mutation, confirming previously undiagnosed HH.[16] Other alleles, such as H63D, are less frequently associated with clinical disease,[13] but may be more commonly found in non–cancer-associated, idiopathic, dilated cardiomyopathy compared with healthy controls.[17]

Given the importance of iron,[8, 12] and the potential genetic involvement[18, 19] in anthracycline-induced cardiac injury, we determined HFE disease-associated allelic frequency in patients considered to be at risk of doxorubicin cardiotoxicity. Awareness of the frequency of the HFE disease allele in children with ALL who are treated with doxorubicin may help clarify the relationship between their genetic predisposition to cardiotoxicity and cardiovascular status. If HFE gene mutations predict late cardiotoxicity, then patients with HFE mutations could be identified at the time of diagnosis and be given cardioprotective agents to minimize their cardiac risk.



Between 2005 and 2007, patients with high-risk ALL treated on Dana-Farber Cancer Institute (DFCI) ALL Consortium protocols from 1991 onward and who had no prior history of disease recurrence were enrolled in this study. Informed consent was obtained from parents if the patient was aged ≤ 18 years and directly from patients aged > 18 years. The total planned cumulative doxorubicin dose for high-risk patients was 300 mg/m2 to 360 mg/m2.

HFE Genotyping

Five mL of peripheral blood was drawn from each patient, with 2.5-mL shipped to the DFCI central testing laboratory for genetic analysis.

Cell Purification

Mononuclear cell fractions were obtained using standard methods. Blood was diluted with an equal volume of serum-free medium and mixed gently. The sample was layered on top of Ficoll (10 mL of Ficoll and 25 mL of diluted blood per 50-mL conical bottom Falcon tube [BD Biosciences, San Jose, Calif]) and then centrifuged for 30 minutes at 1500 revolutions per minute at room temperature. The mononuclear cell layer from the Ficoll:plasma interface was placed into a new 50-mL tube and brought to 5-mL with RPMI medium. The pellet was washed twice, resuspended, and counted in an automated cell counter. Cell concentrations were adjusted to approximately 106/mL. The pellet was frozen in 1-mL aliquots at −80°C for several hours and transferred to liquid nitrogen for long-term storage. Approximately 1.5 × 106 cells were kept unfrozen to generate DNA.

DNA Isolation

Isolated DNA was prepared using NucleoSpin DNA isolation kits (BD Biosciences/Clontech, Mountain View, Calif) in all samples at the time of mononuclear cell fraction isolation. Sample volume was adjusted to 200 μL (1.5 × 106 cells), and 25 μL to 50 μL proteinase K was added. After adding 200 μL to 400 μL of Buffer B3, the mixture was incubated at 70°C for 60 minutes. The sample was vortexed after 210 μL to 420 μL of 96% to 100% ethanol was added, placed in the NucleoSpin column, and centrifuged at 11,000 × g for 1 minute at room temperature. The flow-through was discarded; 500 μL of buffer BW was added to the spin column and centrifuged at 11,000 × g for 1 minute at room temperature. Buffer B5 (600 uL) was added to the spin column and centrifuged at 11,000 × g at room temperature. The flow-through was discarded and the column was recentrifuged at 11,000 × g at room temperature to remove B5 completely. The column was placed in a 1.5-mL tube, eluted with 50 μL to 100 μL of BE buffer (warmed to 70°C) by incubation for 10 minutes, and then centrifuged. Contaminating protein and DNA were measured by ultraviolet spectroscopy. DNA was stored at −80°C.


The 2 most common HFE alleles associated with HH, the cysteine-to-tyrosine substitution at amino acid position 282 (C282Y) and the histidine-to-asparatic acid substitution at amino acid position 63 (H63D), were detected by directly sequencing genomic DNA with a commercial, clinically validated kit (PyroMark HFE Cat #40-0053; Biotage LLC, Charlotte, NC) run on a Pyrosequencer instrument (PSQ HS 96; Biotage) or by Sequenom (Sequenom Inc, San Diego, Calif) and TaqMan (Life Technologies Inc, Carlsbad, Calif) genotyping assays performed at the Harvard Partners Center for Genetics and Genomics.

Genetic Counseling

Before enrolling, patients and families were invited to undergo genetic counseling regarding study implications. Results were disclosed to patients with a homozygous HFE deficiency because this may precede clinically significant iron overload, and retesting in a Clinical Laboratory Improvement Amendments (CLIA)-certified laboratory was advised.

Sample Collection for Biomarkers

The DFCI protocol 91-01 study[20] was conducted before serum cardiac biomarkers were collected. Patients from this cohort were excluded from the cardiac biomarker analysis, but were included in all other analyses.

At the time of enrollment on DFCI protocols 95-01 and 00-01, serum samples were collected at the time of diagnosis, daily for 1 week after the first induction doxorubicin dose, before each subsequent doxorubicin dose, and for 7 days after the final doxorubicin dose.[20] Serum was stored at −70°C until analysis. Hemolyzed samples were excluded. Cardiac troponin-T (cTnT) concentrations were determined centrally using the Elecsys Troponin-T STAT Immunoassay (sensitivity, 0.01 ng/mL; Roche Diagnostics Corporation, Indianapolis, Ind). Concentrations of N-terminal pro-brain natriuretic peptide (NT-proBNP) were measured using an immunoassay (sensitivity, 5 pg/mL; Elecsys immunoanalyzer, Roche Diagnostics). Concentrations of cTnT > 0.01 ng/mL and NT-proBNP concentrations > 150 pg/mL in infants aged < 1 year or > 100 pg/mL in children aged ≥ 1 year at the time of sample collection indicated myocardial injury and cardiomyopathy, respectively.

Cardiac Assessment

Echocardiograms were obtained at the time of diagnosis, after doxorubicin therapy, and every 2 years thereafter at local treatment sites and were centrally remeasured at a single facility by study staff blinded to treatment status.[21] Children were eligible for cardiac follow-up throughout their first continuous complete remission. LV status was assessed with LV end-systolic and end-diastolic dimensions; LV mass; LV end-systolic and end-diastolic posterior wall thicknesses; LV thickness-to-dimension ratio; and LV fractional shortening, an index of LV systolic performance influenced by heart rate, LV preload, LV afterload, and LV contractility.

We standardized echocardiographic measurements with Z-scores,[22] which are the number of standard deviations the measurement is above or below the mean value of a normative population, to adjust for age, body surface area, and growth-related changes. We calculated Z-scores from the difference between LV outcome values in patients and known values in healthy children, divided by the standard deviation of a distribution of values in healthy children.

We calculated the predicted value for each outcome in healthy children with a regression model, using data from 285 healthy children measured in a single center and in the same manner as the study patients.[23, 24] The Z-scores for LV mass, LV end-systolic and end-diastolic posterior wall thicknesses, and LV dimensions were adjusted for body surface area, and Z-scores for LV fractional shortening were adjusted for age at echocardiography.

Statistical Analysis

Means of normally distributed outcome variables (eg, echocardiographic measurements) were compared with Student t tests; means of non-normally distributed variables (eg, age) were compared using Wilcoxon rank sum tests. Differences in the percentages of dichotomous values (eg, mutation carrier) between groups were compared with Fisher exact tests.

For associations with serum cardiac biomarkers, each HFE variant was dichotomized into carriers or noncarriers, and wild-type only. For associations with echocardiographic Z-scores, the 2 loci were combined as a result of a small sample. Carriers were heterozygous or homozygous for either H63D or C282Y, and noncarriers were wild-type only for both. Unadjusted and adjusted logistic regression analyses of multiple abnormalities in NT-proBNP and cTnT concentrations were performed to determine the association with HFE carrier status for both H63D and C282Y both combined and separately. The models were adjusted for treatment including doxorubicin with and without dexrazoxane. All analyses were performed using SAS statistical software (version 9.2; SAS Institute Inc, Cary, NC).


For the 184 high-risk ALL patients, the median time from registration on the original ALL therapy protocol to enrollment on the HFE testing protocol was 7.0 years (range, 1.1 years-17.1 years). Of these, 68 patients (37%) had received doxorubicin alone and 116 (63%) had received dexrazoxane and doxorubicin (Table 1).

Table 1. Patient Characteristics
CharacteristicNo. (%)
  1. Abbreviation: DFCI, Dana-Farber Cancer Institute; HFE, hemochromatosis.

  2. a

    Data were available from only 97 patients.

No. of patients evaluable184
Median time from registration on original DFCI protocol to registration on HFE protocol (range), y7.0 (1.1-17.1)
Male101 (55)
Female83 (45)
Median age at diagnosis (range), y6.3 (<1-17.9)
<10109 (59)
≥1075 (41)
Median age at time of enrollment on DFCI HFE protocol (range), y15.2 (3.1-31.4)
<1044 (24)
≥10140 (76)
Immunophenotype at diagnosis 
T-cell34 (18)
B-cell150 (82)
Median leukocyte count at diagnosis (range), k/µL20.1 (1.3-740.4)
Median serum iron concentration (range), µg/dLa75.0 (15.0-235.0)
Median serum ferritin concentration (range), ng/mLa67.2 (7.21-1998.0)
Doxorubicin68 (37)
Dexrazoxane before doxorubicin116 (63)
Median cumulative doxorubicin dose (range), mg/m2300 (204-420)
Median time from registration on original DFCI protocol to postbaseline echocardiogram (range), y6.1 (1.0-16.1)
No. of echocardiograms 
Time 1 (1-3.99 y from registration)95 (52)
Time 2 (4-6.99 y from registration)53 (29)
Time 3 (≥7 y from registration)47 (26)

Genetic detection was successful for 172 patients for H63D and 179 patients for C282Y. Two patients (1%) were homozygous and 39 (23%) were heterozygous for H63D. One patient (< 1%) was homozygous and 17 patients (9%) were heterozygous for C282Y. Of the 167 patients with results for both tests, 113 (68%) were homozygous wild-type at both loci.

The percentage of patients with ≥ 2 abnormal cTnT concentrations during doxorubicin therapy was higher in C282Y carriers than in noncarriers (31% vs 6%) (Table 2). Multiple elevations in the cTnT concentration were found to be associated with carriers of the C282Y allele (odds ratio [OR], 7.23 [95% confidence interval (95% CI), 1.78-29.4]; P = .006) on univariate analyses. This association remained significant after adjusting for dexrazoxane treatment (adjusted OR, 9.21 [95% CI, 1.11-76.5]; P = .039). No associations were found with NT-proBNP. Neither H63D heterozygosity nor homozygosity was found to be associated with either cardiac biomarker (data not shown).

Table 2. Allelic Variants by Patient Characteristics
 Combined H63D/C282Y (n=167)H63D (n=172)C282Y (n=179)
  1. Abbreviations: cTnT, serum cardiac troponin-T; NT-proBNP, N-terminal pro-brain natriuretic peptide.

  2. a

    Other indicates patients with no abnormal cTnT/NT-proBNP biomarker measurements or <2 abnormal measurements.

 (n=54)(n=113) (n=41)(n=131) (n=18)(n=161) 
Median age at diagnosis (range), y6.8 (<1-17.9)6.1 (<1-17.1).519.8 (1.5-17.6)6.1 (<1-17.9).223.9 (<1-17.6)7.0 (<1-17.6).44
Median age at enrollment (range), y15.2 (4.6-31.4)15.2 (3.1-27.2).5015.8 (6.4-31.4)15.2 (3.1-29.2).2714.3 (4.6-29.2)15.3 (3.1-31.4).61
Male sex, no. (%)33 (61)58 (51).2523 (56)71 (54).8613 (72)85 (53).14
Treatment, no. (%)         
Doxorubicin only20 (37)41 (36).9914 (34)49 (37).859 (50)57 (35).30
Dexrazoxane before doxorubicin34 (63)72 (64) 27 (66)92 (63) 9 (50)104 (65) 
Median cumulative doxorubicin dose (range), mg/m2300 (204-382)300 (240-366).54300 (204-382)300 (240-420).79300 (288-360)300 (204-382).82
No. of evaluable patients3883 2995 13114 
Multiple abnormal cTnT measurements during doxorubicin treatment4 (11)7 (8).741 (3)10 (11).464 (31)7 (6).015
Othera34 (89)76 (92) 28 (97)85 (89) 9 (69)107 (94) 
No. of evaluable patients3882 2994 13113 
Multiple abnormal NT-proBNP measurements during doxorubicin treatment33 (87)72 (88).9925 (86)82 (87).9911 (85)100 (88).65
Othera5 (13)10 (12) 4 (14)12 (13) 2 (15)13 (12) 

At a median of 2.2 years after diagnosis (range, 1.0 years-3.6 years), mean Z-scores for LV mass and end-systolic posterior wall thickness were significantly worse than normal for all children (carriers: mean Z-score, −0.84 [standard error (SE), 0.17; P < .001]; and −4.36 [SE, 0.26; P < .001], respectively; noncarriers: −0.45 [SE, 0.15; P = .006] and −4.06 [SE, 0.17; P < .001], respectively). However, carriers alone also had abnormally low mean Z-scores for LV fractional shortening (−0.71 [SE, 0.25]; P = .008) and end-diastolic posterior wall thickness (−0.68 [SE, 0.25]; P = .011) (Table 3). Although the power to detect such differences was reduced in later follow-up due to the number of echocardiograms available, similar results were obtained and no differences in Z-scores between carriers and noncarriers were detected. The median follow-up for the 114 patients with a postbaseline assessment was 6.1 years (range, 1.0 years-16.1 years).

Table 3. Left Ventricular Structure and Function Z-Scores From Echocardiograms
 Combined H63D/C282Y
Left Ventricular MeasurementHFE Carrier Mean (SE)Carrier Versus Normal PaHFE Noncarrier Mean (SE)Noncarrier Versus Normal PaCarrier Versus Noncarrier Pa
  1. Abbreviations: HFE, hemochromatosis; SE, standard error.

  2. a

    Bold type indicates statistical significance.

Time 1 (1 to 3.99 y)     
End-systolic dimension0.28 (0.17).110.02 (0.21).90.41
End-diastolic dimension−0.09 (0.16).55−0.27 (0.19).16.54
Fractional shortening−0.71 (0.25).008−0.60 (0.30).053.81
Thickness-to-dimension ratio−0.50 (0.33).120.29 (0.31).32.09
Mass−0.84 (0.17)<.001−0.45 (0.15).006.14
End-systolic posterior wall thickness−4.36 (0.26)<.001−4.06 (0.17)<.001.33
End-diastolic posterior wall thickness−0.68 (0.25).011−0.06 (0.22).77.08
Time 2 (4 to 6.99 y)     
End-systolic dimension0.20 (0.34).560.62 (0.23).012.29
End-diastolic dimension−0.43 (0.39).280.25 (0.17).15.08
Fractional shortening−0.93 (0.27).003−1.05 (0.53).054.86
Thickness-to-dimension ratio0.07 (0.31).82−0.63 (0.28).031.11
Mass−0.96 (0.32).009−0.79 (0.21)<.001.64
End-systolic posterior wall thickness−4.63 (0.34)<.001−4.13 (0.21)<.001.20
End-diastolic posterior wall thickness−0.29 (0.39).46−0.61 (0.27).032.49
Time 3 (7 to ≥7 y)     
End-systolic dimension−0.08 (0.27).770.55 (0.27).051.18
End-diastolic dimension−0.20 (0.38).620.24 (0.21).26.29
Fractional shortening−0.27 (0.45).56−0.59 (0.27).033.51
Thickness-to-dimension ratio−1.13 (0.52).045−0.94 (0.19)<.001.66
Mass−1.15 (0.59).092−1.04 (0.32).004.87
End-systolic posterior wall thickness−4.16 (0.86).005−4.20 (0.26)<.001.95
End-diastolic posterior wall thickness−0.77 (0.56).20−1.08 (0.22)<.001.54


In the current study of survivors of childhood high-risk ALL, heterozygosity for C282Y was associated with multiple elevations in cTnT concentrations after controlling for dexrazoxane treatment. Furthermore, compared with a normal population, patients with the C282Y and/or H63D allelic variants had significantly lower LV function, LV mass, and wall thickness 2 years after diagnosis.

The HFE prevalence in the current study is similar to that of the US population, in which the estimated prevalence for C282Y-H63D heterozygosity is between 1.5% and 2.5%.[25] For the C282Y mutation, the prevalence is 7% to 9% for heterozygotes and 0.12% to 0.5% for homozygotes.[25] For the H63D mutation, the prevalence is 20% to 23% for the heterozygotes and 1.5% to 2.4% for the homozygotes.[25]

The results of both in vitro and animal studies have suggested that iron-loaded tissues enhance anthracycline cardiotoxicity, which increases concentrations of markers of cellular damage.[12, 26] HFE gene mutations predispose rodents to doxorubicin-induced cardiotoxicity.[26] Mice with a targeted mutation of the HFE gene (homozygotic HFE / mice) were found to have iron overload in multiple organs, including the heart, compared with wild-type controls.[26] In addition, when treated with doxorubicin, HFE / mice were found to have concentrations of serum markers for acute cardiac injury, mitochondrial damage, myofibril degeneration, and mortality that were significantly higher than those in wild-type mice. In HFE heterozygotic mice (HFE+/), chronic doxorubicin administration caused mitochondrial degeneration and increased mortality rates, although not to the same extent as in HFE / animals.[26] This result suggests that homozygous or heterozygous mutations of HFE or perhaps even of other genes linked to increased iron stores and HH may increase susceptibility to doxorubicin cardiotoxicity.

The magnitude of the Z-score differences in the current study, although small relative to those that guide the daily clinical decisions made by cardiologists, are consistent with findings in pediatric cancer survivors treated with doxorubicin,[4-6] who years later have increased rates of congestive heart failure (CHF) and cardiac mortality compared with controls.[27] Although there were no significant differences noted between carriers and noncarriers in the current study, the Z-scores are characteristic of anthracycline-associated cardiomyopathy in that acute evidence of injury is often followed by normalization of ventricular size and function, presumably by ventricular remodeling (cellular hypertrophy compensating for cell loss). Nonetheless, late evidence of abnormal ventricular structure with reduced LV wall thickness is a frequent outcome.[4] Furthermore, the risk of anthracycline-related CHF may be modified by the presence of HFE variants, as shown in anthracycline-treated survivors of hematopoietic cell transplantation.[28]

Cascales et al retrospectively evaluated cardiac iron, cardiac events, and HFE genotypes (C282Y and H63D) in 97 consecutive autopsy results from patients with solid and hematological cancers, 48 of whom had been treated with anthracyclines and 49 who received no chemotherapy (n = 25) or nonanthracycline chemotherapy (n = 24).[29] Patients treated with cumulative anthracycline doses > 200 mg/m2 were found to have higher heart iron concentrations (490 μg/g dry weight vs 240 μg/g dry weight; P = .01), independently of liver iron load or transfusion history, compared with controls. Mutations in HFE were associated with higher iron deposits in the heart tissue, but not with global cardiac events or CHF. Multivariate linear regression demonstrated that both HFE genotypes and anthracyclines contributed to cardiac iron concentrations (ρ = 0.284), a result suggesting cardiac iron accumulation modulation by both anthracyclines and HFE and thereby supporting the significance of HFE status as a predictor of anthracycline cardiotoxicity.[29]

Myocardial injury (elevated cTnT) during doxorubicin therapy in patients with C282Y alleles puts them at a greater risk of subsequently reduced LV wall thickness and mass due to loss of cardiomyocytes during therapy.

In the current study, carriers were found to have significantly reduced LV function 2 years after the diagnosis of ALL compared with healthy children, a finding that was only marginally observed when ALL noncarriers were compared with healthy children. This suggests that the remaining cardiomyocytes are adversely affected by the combination of HFE carrier status and doxorubicin therapy, going beyond cardiomyocyte demise during therapy (elevated cTnT concentrations), resulting in late reduced LV wall thickness and mass being the sole type of late doxorubicin cardiotoxicity. Unhealthy residual myocardium in long-term ALL survivors treated with doxorubicin who are also carriers of HFE alleles is supported by a reduced LV thickness-to-dimension ratio, a trend consistent with early pathologic LV remodeling in this population. Such an association has been described in HFE carriers without cancer with idiopathic dilated cardiomyopathy.[13]

Patients with C282Y homozygosity are reported to more frequently have genetic variations in antioxidant enzymes and mitochondrial DNA, which may increase their susceptibility to myocardial iron-induced oxidative stress and dysfunction, respectively.[30, 31]

Dexrazoxane iron chelation therapy may prevent the enhanced doxorubicin-mediated damage in an iron-overloaded heart.[32] Children with high-risk ALL who are treated with dexrazoxane before doxorubicin administration are reported to have fewer and milder subclinical biomarker and echocardiographic signs of cardiotoxicity.[32-38]

Identifying all patients at high risk of anthracycline-induced cardiotoxicity is difficult because factors known to increase this risk do not identify many patients who subsequently develop cardiomyopathy.[6, 8] Although generalized population screening for HFE mutations is not recommended, case-finding by using HFE screening in high-risk groups to detect affected individuals might be beneficial.[39] Screening for HFE mutations common in children with newly diagnosed high-risk ALL who are to be treated with anthracyclines might inform treatment decisions regarding chemotherapy, cardioprotectant therapy, transfusions, and magnetic resonance imaging to determine early myocardial iron overload.[13, 40]


Supported in part by grants from the National Institutes of Health (HL072705, HL078522, HL053392, CA127642, CA068484, HD052104, AI50274, HD052102, HL087708, HL079233, HL004537, HL087000, HL007188, HL094100, HL095127, and HD80002), the Children's Cardiomyopathy Foundation, the Women's Cancer Association of the University of Miami, the Lance Armstrong Foundation, the STOP Children's Cancer Foundation, the Scott Howard Fund, and the Michael Garil Fund.


Dr. Kutok is currently a paid employee of Infinity Pharmaceuticals Inc but was not at the time the current study was conducted. He owns stock in Infinity Pharmaceuticals. Dr. Fleming has acted as a consultant for Millennium Pharmaceuticals, Epizyme, and Bayer and has received grants from Alnylam Pharmaceuticals.