Cardiovascular Family History Increases Risk for Late-Onset Adverse Cardiovascular Outcomes in Childhood Cancer Survivors: A St. Jude Lifetime Cohort Report

Children diagnosed with a malignancy have 5-year survival rates greater than 83% as a result of therapeutic advances (1). However, survivors have well-documented treatment-related late effects (2, 3). Cardiovascular disease is one of the most frequent causes of non–cancer-related morbidity and mortality (4–10). Compared with the general population, survivors have higher rates of: (i) cardiovascular diseases, specifically, myocardial infarction, stroke, and heart failure, and (ii) cardiovascular risk factors, including, hypertension, obesity, diabetes mellitus, and dyslipidemia (5, 7, 9, 11, 12). Previous work identified that survivors were seven times more likely than the general U.S. population to die from cardiovascular-related events (5). Known risk factors for the development of cardiovascular disease in survivors of childhood cancer include cumulative dose of anthracycline chemotherapy and chest-directed radiotherapy, young age at anthracycline exposure, female sex, and black race (5–8, 13). More recently, attention has been given to the management of cardiovascular risk factors in the survivor population, as hypertension, diabetes, and obesity have been shown to potentiate risk of cardiovascular events among survivors (9, 14, 15).

In the general population, Hunt and colleagues first described cardiovascular family history as a risk factor for cardiovascular disease in a large cohort study of >90,000 Utah school children and their families. The importance of cardiovascular family history has since been replicated in other large series, such as the Newcastle Family History Study and the Framingham Offspring Study (16–18). While much attention has been given to the significant cardiovascular risk conferred by a family history of early or premature cardiovascular disease, more recent studies show similar risk contribution from the presence of any first-degree family history of cardiovascular disease (19, 20). The occurrence of hypertension and diabetes in the general population can be similarly predicted by a family history of these conditions (21, 22). However, the independent contribution of cardiovascular family history toward the development of cardiovascular disease or traditional cardiovascular risk factors among survivors of childhood cancer is not known. It might have been predicted that in patients exposed to cardiotoxic therapies, the exposure would be so overwhelming that family history may not contribute to outcomes. Therefore, it is important to examine family history's contribution to cardiovascular risk in both survivors who have been exposed to cardiotoxic therapy as well as those who have not.

Efforts to mitigate modifiable risk factors for cardiovascular disease in the survivor population may be improved with the addition of knowledge of cardiovascular family history. Therefore, we utilized the St. Jude Lifetime Cohort Study (SJLIFE) to examine the contribution of family history of cardiovascular disease and cardiovascular risk factors to the development of these same conditions among survivors of childhood cancer exposed and unexposed to cardiotoxic therapies.

Eligible participants were identified from the SJLIFE cohort, designed to characterize health outcomes among survivors of childhood cancer as they age (23). Survivors eligible for this study were ≥18 years of age at evaluation, treated for childhood cancer at St. Jude Children's Research Hospital (Memphis, TN) between 1962 and 2007, ≥10 years postdiagnosis, and without a history of congenital heart disease. Females were not pregnant and ≥3 months postpartum. Study documents, materials, and methods were approved by the institutional review board. Participants provided written informed consent prior to participation.

Cancer diagnosis and treatment information were abstracted from medical records including cumulative doses of specific chemotherapy agents and radiotherapy. Total anthracycline dose (mg/m²) was the sum of doxorubicin, daunorubicin, epirubicin, idarubicin, and mitoxantrone in doxorubicin equivalent doses (24, 25). Radiotherapy records were centrally reviewed for the receipt of chest radiotherapy. In this analysis, participants were grouped as being exposed or unexposed to cardiotoxic therapy (anthracycline chemotherapy and/or chest radiotherapy; Table 1).

A provider-administered family history questionnaire evaluated the presence and age at onset of cardiovascular disease or cardiovascular risk factors among all first- and second-degree relatives of each participant. The cardiopulmonary family history questionnaire was based on the Multi-Ethnic Study of Atherosclerosis (MESA) Family History Form (Supplementary Table S1A). MESA investigated the prevalence, correlates, and progression of subclinical cardiovascular disease. Family history elements included in the analysis were first- and second-degree family history of: cardiovascular disease (myocardial infraction, stroke, or heart failure), atherosclerotic disease (myocardial infraction or stroke), early atherosclerotic disease (myocardial infraction or stroke occurring before age of 55 years in men and before age of 65 years in women; ref. 24), and cardiovascular risk factors (limited to hypertension or diabetes). Family history of hyperlipidemia was not present in the MESA study or in this assessment, given the poor reliability of self-reporting of this diagnosis. Each relative's sex and age at last birthday or age at death were recorded.

At baseline and subsequent follow-up evaluations, participants completed a multi-item questionnaire that included age at onset of cardiac events, as well as comprehensive clinical assessment, including echocardiography, to classify cardiovascular outcomes using a modified version of the NCI's Common Terminology Criteria for Adverse Events (CTCAE, version 4.03; refs. 25, 26; Supplementary Table S2A). We included three cardiovascular disease outcomes (myocardial infarction, stroke, and heart failure) and two cardiovascular risk factors. CTCAE events are graded 1–5: mild (grade 1), moderate (grade 2), severe (grade 3), life-threatening or disabling (grade 4), and fatal (grade 5). We included stroke outcomes of grade 2 (moderate neurologic deficit), grade 3 (severe neurologic deficit), and grade 4 (life-threatening consequences requiring urgent intervention). We included heart failure outcomes of grade 2 [resting ejection fraction (EF) between 40% and 50% or 10% and 19% absolute drop from baseline; asymptomatic], grade 3 (resting EF between 20% and 39% or >20% absolute drop from baseline; medication initiated), and grade 4 (resting EF < 20%; refractory heart failure requiring intravenous medication or surgical intervention). For myocardial infarction, we included outcomes of grade 3 (abnormal cardiac enzymes with ECG evidence of infarction) and grade 4 (life-threatening myocardial infarction with hemodynamic instability). We excluded grade 2 myocardial infarction as it describes mildly abnormal cardiac enzymes without ECG evidence of ischemia, rather than myocardial infarction. We included grades 2–4 hypertension and diabetes.

Descriptive statistics characterized the study population, and compared survivors exposed and not exposed to anthracyclines and/or chest-directed radiotherapy with t tests or χ² statistics as appropriate. Logistic regression analyses, stratified by exposure to cardiotoxic treatment, were used to examine univariable associations between family history and cardiovascular outcomes. Multivariable risk models were created based upon statistically significant associations (P < 0.05) identified from univariable analysis and were adjusted for sex, race/ethnicity, and current age, as well as an exposure variable combining anthracycline dose and chest radiation exposure among exposed survivors (1–199, 200–349, and 350+ mg/m² anthracycline alone, and chest radiotherapy, no anthracycline, chest radiotherapy, any anthracycline). Receiver operating curve analysis was conducted to evaluate the AUC defined by models including family history variables as added to models evaluating treatment exposure (27). SAS version 9.4 was used for all analysis. Cumulative burden curves were constructed among individuals with and without family history of cardiovascular disease or cardiovascular risk factors, evaluating the development of cardiovascular disease or risk factor over time (3).

Among 1,369 eligible survivors, 67 declined participation, 24 completed a survey only (did not return for clinical evaluation of cardiac outcomes), and 18 were lost to follow-up, resulting in enrollment of 1,260 participants (Fig. 1; ref. 28). Survivors exposed (n = 824) and unexposed (n = 436) to cardiotoxic therapy (anthracycline chemotherapy and/or chest-directed radiotherapy) did not differ by median age (range): 36 (19–62) vs. 34 (18–66) years, P = 0.19; sex: 53.2% vs. 47.3% male, P = 0.05; race/ethnicity: 84.8% vs. 81.0% non-Hispanic White, P = 0.33; or time since diagnosis: P = 0.34. Survivors exposed to cardiotoxic therapy were older at diagnosis (P < 0.0001) and were more likely to be survivors of lymphoma or bone tumors than those who were not exposed (Table 1).

Among survivors exposed to cardiotoxic therapy, 7% reported a first-degree relative with heart failure, 19% with myocardial infarction, 11% with stroke, 26% with any atherosclerotic disease (myocardial infarction and/or stroke), 62% with hypertension, and 31% with diabetes (Table 2). The proportions of survivors with a family history of cardiovascular disease or cardiovascular risk factors were similar between the exposed and unexposed groups. A first- or second-degree relative with a history of cardiovascular disease (myocardial infarction, stroke, or heart failure) was present in 73.9% of survivors exposed to cardiotoxic therapies and in 75.9% of unexposed participants. A first- or second-degree relative with a history of cardiovascular risk factors (diabetes or hypertension) was present in 86.3% of survivors exposed to cardiotoxic therapies and in 88.5% of unexposed participants (Supplementary Tables S3A and S4).

At last follow-up, 18% of exposed survivors had developed heart failure, 9% myocardial infarction, 3% stroke, 11% any atherosclerotic disease (myocardial infarction and/or stroke), 30% hypertension, and 9% diabetes. All cardiovascular outcomes were developed greater than 2 years from cancer diagnosis. Upon univariable analysis, the development of heart failure among participants who were exposed to cardiotoxic therapy was associated with first-degree family history of cardiovascular disease [RR, 1.5; 95% confidence interval (CI), 1.1–2.0], atherosclerotic disease (RR, 1.5; 95% CI, 1.1–2.0), early atherosclerotic disease (RR, 1.4; 95% CI, 1.0–1.9), any cardiovascular risk factor (RR, 1.5; 95% CI, 1.1–2.1), or first-degree family history of cardiovascular disease or cardiovascular risk factors (RR, 1.6; 95% CI, 1.1–2.3; Table 3). There was no association between family history of heart failure and development of heart failure in survivors. Furthermore, first-degree family history of any cardiovascular disease or cardiovascular risk factors was associated with an increased risk for hypertension and diabetes in exposed survivors. While few survivors unexposed to cardiotoxic therapies developed heart failure (n = 15, 3%), myocardial infarction (n = 14, 3%), stroke (n = 15, 3%), any atherosclerotic disease (n = 30, 7%), hypertension (n = 140, 32%), or diabetes (n = 43, 10%), unexposed survivors with a first-degree relative with heart failure (RR, 9.4; 95% CI, 3.6–24.5) and with a first-degree relative with cardiovascular disease (RR, 3.3; 95% CI, 1.2–9.0) were at increased risk for heart failure. Among survivors not exposed to cardiotoxic therapy, hypertension was predicted by family history of hypertension (RR, 1.6; 95% CI, 1.1–2.2), and diabetes was predicted by family history of diabetes (RR, 2.6; 95% CI, 1.5–4.5). No differential effect based on high- or low-dose cardiotoxic exposure was observed for development of heart failure or atherosclerotic disease (Supplementary Tables S5A–S8A).

Multivariable analysis among patients exposed to cardiotoxic therapies, adjusting for sex, race/ethnicity, age at evaluation, and treatment, demonstrated that a family history of atherosclerotic disease in a first-degree relative was independently associated with an increased risk for treatment-related heart failure (RR, 1.38; 95% CI, 1.01–1.88; P = 0.04; Table 4; Supplementary Table S9A). Other family history elements showed a similar association with heart failure, but did not achieve statistical significance. Having a first-degree relative with hypertension (RR, 1.55; 95% CI, 1.26–1.92; P < 0.0001), cardiovascular disease (RR, 1.30; 95% CI, 1.06–1.59; P = 0.0103), cardiovascular risk factors (RR, 1.58; 95% CI, 1.24–2.01; P = 0.0002), and cardiovascular disease or cardiovascular risk factors (RR, 1.71; 95% CI, 1.27–2.28; P = 0.0003) increased risk for hypertension (Table 5). Among survivors not exposed to cardiotoxic treatments, family history was not associated with the development of heart failure or hypertension in multivariable analyses (Supplementary Tables S10A and S11A). Among exposed and unexposed survivors, a family history of diabetes and other cardiovascular diseases/risk factors in first-degree relatives was associated with the development of diabetes (Supplementary Tables S12A and S13A).

The cumulative burden of cardiovascular outcomes and risk factors was greater among those with a family history of cardiovascular disease or risk factor [0.97 (95% CI, 0.88–1.06) vs. 0.59 (95% CI, 0.48–0.71) at 45 years from diagnosis and 0.91 (95% CI, 0.83–1.00) vs. 0.58 (95% CI, 0.46–0.70) at attained age of 50 years; Fig. 2). Receiver operating curve analysis for models of treatment-exposed individuals identified that addition of family history of hypertension improved AUC for prediction of hypertension. Adding remaining model elements did not improve AUC (Supplementary Table S14A).

Long-term survivors of childhood cancer are at increased risk of cardiovascular disease and cardiac-specific mortality due to exposure to therapies with known cardiovascular effects (anthracycline chemotherapy and chest-directed radiotherapy; refs. 7, 14, 29, 30). Our study demonstrated, in a population of survivors of childhood cancer exposed to cardiotoxic therapies, yet still relatively young for development of heart disease, that: (i) development of heart failure is independently associated with a history of atherosclerotic disease (myocardial infarction and stroke) in a first-degree relative and (ii) development of hypertension is independently associated with a history of hypertension or cardiovascular disease in a first-degree relative. The results also suggest, although the number of cardiac events was low, that an association exists between family history of cardiovascular diseases/risk factors and the development of heart failure and hypertension among survivors not exposed to cardiotoxic therapies, mirroring the known hereditary patterns of cardiovascular risk in the general population. Thus, in this analysis, family history of cardiovascular diseases/risk factors was an independent predictor of cardiovascular outcomes in survivors both exposed and unexposed to cardiotoxic therapies. This is important because family history remains predictive of cardiovascular outcomes even among survivors exposed to highly cardiotoxic treatment, showing that the effect of family history is independent and not overwhelmed by the receipt of cardiotoxic therapies. By identifying that family history is an independent predictor of heart failure, even in the setting of exposure to cardiotoxic therapy, these findings highlight the importance of obtaining and updating the cardiac family history among survivors of childhood cancer at any age. With the additive knowledge of cardiovascular family history in the survivor population, more attention can be paid to the early diagnosis and treatment of cardiovascular conditions, such as hypertension and heart failure, to mitigate poor outcomes commonly observed.

It is important to know whether the prevalence of family history of cardiovascular disease among survivors of childhood cancer is similar to that of the general population. The MESA questionnaire was utilized in this study to determine family history of cardiovascular disease. As such, this sample's prevalence of and risk relative to cardiovascular disease could be compared with those identified by previous studies from MESA data, which cites a 36% prevalence of family history of cardiovascular disease and a 16% prevalence of early cardiovascular disease family history (20). These prevalence values are similar to those found in our sample (exposed survivors having 28.2% family history and 17.1% early family history and unexposed survivors having 31.4% family history and 18.4% early family history), suggesting that the family history of survivors of childhood cancer is not unlike other populations.

Surveillance parameters for adult survivors of childhood cancer are outlined in the Children's Oncology Group (COG) Long-Term Follow-up Guidelines (31). Although these guidelines include recommendations for screening and follow-up assessment of cardiovascular disease and cardiovascular risk factors among recipients of cardiotoxic therapies (9), no recommendations regarding the ascertainment of cardiovascular family history are provided. The conclusions defined by this study provide evidence for the importance of obtaining the history of myocardial infarction, stroke, heart failure, hypertension, and diabetes among first-degree relatives.

This is important, given the known cardiovascular risk imparted by family history of cardiovascular disease and premature cardiovascular disease in the general population, resulting in the American Academy of Family Physicians and American Academy of Pediatrics recommending ascertainment of family history of cardiovascular disease events (coronary death, myocardial infarction, stroke, or hospitalized coronary insufficiency) among first- and second-degree relatives (24, 32, 33). These recommendations are informed by the Framingham Offspring Study, which included two-generation family history data collected between 1948 and 2001 of >5,000 offspring of participants of the original Framingham Heart Study. Offspring participants with at least one parent with premature cardiovascular disease (onset age <55 years in father or <65 years in mother) were more than twice as likely to develop cardiovascular events (24). Similar findings have been replicated in larger, more contemporary cohorts, including the Physicians' Health Study and the Women's Health Study, informing this definition of premature family history of cardiovascular disease as the gold-standard for cardiovascular disease family history assessment (19). Further studies of cardiovascular family history, however, provide evidence that a simplified definition (any first-degree relative with coronary artery disease) may be similarly predictive of familial cardiovascular risk (17, 20). The complex care of the childhood cancer survivors includes multi-system surveillance for cancer recurrence and treatment-related chronic conditions, and may neglect the complete ascertainment of familial cardiovascular disease risk. Given the cardiovascular disease risk attributed to the presence of cardiovascular disease in first-degree relatives of this relatively young population, coupled with knowledge of the high risk for early onset of cardiovascular disease in the survivor population, childhood cancer survivors should have early and serial cardiovascular family history assessment (7, 34).

Consistent with the association of cardiovascular disease with a family history of cardiovascular disease in the general population, we identified that among survivors, cardiovascular risk factors, specifically, hypertension and diabetes, were strongly associated with family history of these diagnoses (21, 22, 35). Furthermore, in the general population, longitudinal studies have established bidirectional associations between the presence of cardiovascular disease and hypertension/diabetes within personal and family histories (21, 22, 35). As described previously, COG guidelines include assessment of hypertension and diabetes among childhood cancer survivors, given the known associations between cardiotoxic therapies and these outcomes. Our results show associations of family history of both cardiovascular disease and cardiovascular risk factors with the development of hypertension and diabetes. The information provided by this study provides additional evidence regarding the importance of prevention and successful treatment of hypertension and diabetes, now more so in survivors with a family history of cardiovascular diseases.

The study's strengths include direct assessment of cardiac outcomes in a large population of long-term survivors and interviewer-directed family history assessment, with detailed records of cancer therapy and relatively long interval of longitudinal evaluation. The main limitation of this study, however, is the inability to validate survivor-reported family history information. Previous studies have evaluated the reliability of self-reported family history. An evaluation of the Framingham Offspring Study, for example, found high positive predictive values (>75%) for self-reported family history of hypertension and diabetes, but poorer such values for family history of premature heart attack or stroke (28% and 43%, respectively; ref. 36). Given these limitations, future investigations should define the accuracy and predictive value of this family history data through medical record validation. In addition, future studies can utilize genetic risk scoring data that have been collected within this sample to strengthen evaluation of familial-based cardiovascular risk in this population. Another potential limitation is that all participants were treated for their childhood cancer at a single institution. Although previous evaluation of SJLIFE patients has shown participants to have similar exposures to that of other children treated in North America during this time period, these results may not be generalizable to all survivors (3, 23, 37). In addition, the study had a limited number of individuals in the unexposed group (n = 430), compared with the exposed group (n = 824), which may have affected the ability to achieve appropriate statistical power among unexposed individuals. A predominance of non-Hispanic White race/ethnicity is a known limitation of the dataset. Given the known effects of race on cardiovascular outcomes, future survivor datasets should be powered to evaluate these differences.

In summary, among survivors who received anthracycline chemotherapy or chest-directed radiotherapy, a history of myocardial infarction and stroke in a first-degree relative was independently associated with the development of heart failure. In addition, the development of hypertension among exposed survivors was predicted by a history of cardiovascular disease and cardiovascular risk factors in a first-degree relative. The results also suggest an elevated risk of heart failure, hypertension, and diabetes in survivors with a family history of cardiovascular disease and cardiovascular risk factors and no exposure to cardiotoxic therapies, as would be expected in the general population. These findings underscore the importance of cardiovascular family history screening among survivors of childhood cancer, which can start upon diagnosis of a malignancy and continue through the long-term follow-up period. With more robust collection of cardiovascular family history data and cardiovascular genetic risk factors, those caring for the survivor population may be able to better identify and intervene upon cardiovascular disease, attenuating the cardiovascular morbidity and mortality known to this population.

K.K. Ness reports grants from NIH during the conduct of the study. R.V. Luepker reports grants from NIH-government during the conduct of the study. R.E. Partin reports grants from NCI and Cancer Center Support (CORE) during the conduct of the study. R.M. Howell reports grants from St. Jude Research Hospital (research group has received funding for dosimetry for the SJLIFE cohort) during the conduct of the study. M.M. Hudson reports grants from NCI (U01CA195547) during the conduct of the study. L.L. Robison reports grants from NIH during the conduct of the study. G.T. Armstrong reports grants from NCI during the conduct of the study. No disclosures were reported by the other authors.

J.F. Goldberg: Conceptualization, resources, investigation, methodology, writing–original draft, writing–review and editing. K.K. Ness: Conceptualization, resources, data curation, formal analysis, validation, investigation, writing–review and editing. X. Chi: Data curation, formal analysis. A.K. Santucci: Funding acquisition, investigation, methodology, writing–review and editing. J.C. Plana: Conceptualization, funding acquisition, investigation, methodology, writing–review and editing. V.M. Joshi: Conceptualization, resources, methodology, writing–review and editing. R.V. Luepker: Conceptualization, resources, supervision, investigation, methodology, writing–review and editing. J.-B. Durand: Conceptualization, funding acquisition, methodology, writing–review and editing. R.E. Partin: Conceptualization, funding acquisition, methodology, writing–review and editing. R.M. Howell: Conceptualization, funding acquisition, methodology, writing–review and editing. C.L. Wilson: Conceptualization, funding acquisition, investigation, methodology, writing–review and editing. J.A. Towbin: Conceptualization, resources, supervision, investigation, methodology. J.L. Jefferies: Conceptualization, resources, supervision, investigation, methodology, writing–review and editing. D.K. Srivastava: Conceptualization, resources, data curation, formal analysis, investigation, visualization, methodology, writing–review and editing. M.M. Hudson: Conceptualization, resources, data curation, supervision, funding acquisition, investigation, methodology, project administration, writing–review and editing. L.L. Robison: Conceptualization, resources, data curation, supervision, funding acquisition, investigation, methodology, project administration, writing–review and editing. G.T. Armstrong: Conceptualization, resources, data curation, supervision, funding acquisition, investigation, methodology, project administration, writing–review and editing.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.