Cardiovascular disease risk among female breast cancer survivors: findings from the National Health Interview Survey (2019–2022)

Introduction

According to the American Cancer Society, breast cancer (BC) is the second most prevalent cancer among women in the United States (US) after skin cancer. One in eight women (13%) will be diagnosed with BC in her lifetime (1). The median age of patients diagnosed with localized BC is 63 years old, and the current 5-year survival rate is 91% (2, 3).

Cardiovascular disease (CVD) is also a leading cause of morbidity and mortality among women. CVD includes major diagnoses such as coronary artery disease, heart failure, myocardial infarction, as well as other conditions such as angina and preclinical heart disease (4, 33). It is estimated that approximately 44% of women have some form of CVD (5, 6).

BC and CVD share many clinically relevant risk factors and comorbidities. Socio-demographic factors like sex, race/ethnicity, and age, are important non-modifiable risk for both BC and CVD, (7) while modifiable lifestyle risk factors include physical activity levels, tobacco use, nutrition and health care access among a host of other factors (8, 9).

Breast cancer incidence rates have increased among all racial/ethnic groups between 2012 and 2021 but racial/ethnic minority, BC populations continue to shoulder a disproportionate burden of both BC and CVD risk (1, 35). The black-white gap in cancer mortality has remained stable. Black women experience a 40% higher rate of BC mortality compared to white women (10, 11). Approximately 23% of African American women with BC are diagnosed at younger than 50 years old, compared with 16% of white women (12). American Indians and Alaska Natives have lower BC incidence rates than white women but experience higher mortality as well. There are parallel differences regarding the outcomes of CVD treatment by race/ethnicity. Black BC survivors experience higher cardiotoxicity and earlier onset of CVD symptoms compared to white BC survivors and are more likely to die prematurely (13, 14). Comorbid conditions such as hypertension, diabetes, hypercholesterolemia, and hyperlipidemia exacerbate the risk of CVD among BC survivors (15). In addition to hypertension, smoking is a particularly high-risk factor in the development of CVD, partly due to its effects on raising levels of low-density lipoprotein cholesterol and endothelial dysfunction and accelerating the progression of atherosclerosis (16).

Cancer treatments may also create cardiotoxicity during active treatment, and may increase incident CVD, exacerbate underlying subclinical CVD, and increase long term CVD risk (17). Radiotherapy has been linked with heart failure and myocardial infarction (18). Chemotherapies such as methotrexate or fluorouracil have also been shown to increase the risk of congestive heart failure (18), while the use of anthracyclines or fluoropyrimidines can lead to endothelial dysfunction and severe cardiac damage. Radiotherapy increases the risk of coronary artery disease due to endothelial disruption, and resultant hypertension (19).

Social determinants of health (SDoH) frameworks for understanding and addressing CVD in BC

The Social Ecological Model (SEM) (37) and the Integrated Conceptual Framework for Understanding and Addressing Social Determinants to Advance Cancer Health Equity (20) provide a conceptual framework for understanding the complex relationship between health behaviors and individual, interpersonal, organizational, societal subsystems. In this observational, cross-sectional study we use these frameworks to inform the selection of the individual-level and community-level variables, publicly available in the National Health Interview Survey (NHIS). We included these variables in bivariate and multivariate modeling of the associations between BC and CVD diagnosis in a national sample of female BC survivors.

The individual level variables include biological and clinical factors that influence CVD risk such as comorbid health conditions, sociodemographic characteristics and preventive health behaviors. Community level factors include measures of access to routine health care and health insurance that impact BC survivors' ability to adequately mitigate CVD risk and cancer recurrence. BC survivors need long term access to follow-up care, routine CVD screening and support to manage obesity, sedentary lifestyles, and poor eating habits. As the population of women with BC ages and as long-term survivorship increases in response to improvements and access to advanced treatment options, the likelihood of CVD comorbidity in BC will rise as well. We need to ensure that implementation of current survivorship recommendations and clinical guidelines mitigate current gaps in patient education and provider standard of care in the immediate post-treatment period and beyond existing 5-year guidelines (21–23).

Methods

Data source

This study utilized data from the NHIS, an ongoing cross-sectional household survey conducted by the National Center for Health Statistics (NCHS). The NHIS has continuously monitored the health of the U.S. civilian, non-institutionalized population since 1957 using a multistage area probability sampling design to generate nationally representative estimates. The survey excludes individuals in long-term care facilities, correctional institutions, and active-duty military personnel in barracks. In 2019, NHIS implemented a comprehensive survey redesign. Therefore, survey data from 2019 to 2022 are most comparable for this analysis.

Study sample

This study included 2,745 women with BC who responded to NHIS in four cohorts from 2019 through 2022. No IRB approval was required to analyze this public dataset.

Outcome variables

Primary outcome: self-reported CVD diagnosis. Using the NHIS self-reported data, the presence of a CVD diagnosis was operationalized as a composite variable that was dichotomized based on whether a respondent reported that a physician confirmed a positive diagnosis of any one of four CVD-related conditions: angina, stroke, myocardial infarction, and coronary artery disease (24).

Select exposure variables

The exposure variables included in the analysis reflect important individual and community level CVD and BC risk factors indicated by the SDoH frameworks undergirding this work. Individual level variables selected from the NHIS survey modules included patient socio-demographics (age, insurance status, employment status), marital status, race, education; number of years since BC diagnosis); preventive health behavior (had doctor visit in last 12 months; delay in medical care, physical activity, smoking history); and health status and comorbid health conditions (BMI, diabetes, hypertension, high cholesterol, self-reported general health status) (Table 1). The original response category for most variables was dichotomous. Four variables (age, marital status; number of years since BC diagnosis; and had a doctor visit in last 12 months) were re-coded into fewer response categories for ease of analysis.

Table 1. National Health Interview Survey (NHIS) modules used in 2019–2022 surveys.

Statistical analysis

We first examined the relationship between BC and CVD from 0 to 10+ years in this sample by comparing the relative prevalence of CVD in five pooled survivor cohorts. This simple trend analysis was used to visualize the relative prevalence of CVD in five survivorship cohorts (< 1 year; 1–<5 years; 5–<7 years; 7–<10 years; 10+ years) in the years 2019, 2020, 2021, and 2022.

Descriptive summaries were then performed for the sociodemographic, behavioral and clinical variables with CVD as the primary outcome. The Wald Chi-square test was used to evaluate the correlations between a self-reported CVD diagnosis and BC survivors' socio-demographic characteristics, clinical comorbidities, SDoH, health-seeking behavior, and self-reported general health status.

To efficiently reduce the number of variables considered in final model and given the relative sample size in each BC group, the decision was made to assess the relative odds of CVD in BC survivors using a forward stepwise logistic regression model. Comorbidities, lifestyle risk and preventive care behaviors were evaluated for multicollinearity and then added to the model after first analyzing Wald Chi-square P-value, and then retained or removed in turn, resulting in a more parsimonious model with the most relevant predictors retained.

Results

CVD prevalence in BC survivors

The final sample included 2,495 women with self-reported BC diagnosis. A total of 250 surveys were removed for missing more than 15% of the variables of interest (Figure 1). The prevalence of CVD among women with BC is highest among those surviving 10 or more years (n = 1,359) and least among those diagnosed 1– < 5 years prior (n = 544). The prevalence of CVD in the cohort of women diagnosed with BC less than 1 year is 16.4%. The prevalence of CVD decreases to 10.5% between 1 and < 5 years after diagnosis and 11.0% during years 5– < 7 after diagnosis. However, there is a notable increase in relative number of BC survivors with CVD 7– < 10 years after diagnosis (13.8%) and even higher after 10 years (19.1%). The higher prevalence of BC survivors with CVD in these survivorship cohorts reflects the expected increased risk of CVD conditions after initial treatment up to 1 year as well as the increased risk of CVD in much older, longer surviving BC patients (Figure 2). However, the spike in the CVD prevalence among BC survivors after 7 years survivorship signals an opportunity to determine the extent to which the increase of CVD among women with BC is independently apportioned to aging, years of survivorship, BC treatment modality, other well-established lifestyle and preventive behaviors; or the concurrent relaxing of enhanced CVD screening guidelines 5 years after BC treatment ends.

Sociodemographic factors and CVD prevalence in women with BC

The cross-sectional association between select sociodemographic variables and CVD in BC survivors was assessed using a Wald Chi-square test (Table 2). Of the 17 variables, 13 variables were found to be significant at α = 0.05. The unique contributions of insurance status; doctor visit in last 12 months; education; and race were not statistically significant between women with CVD and those without CVD and these four variables were not included in the logistic model. All other sociodemographic variables were found to be statistically significant. We determined that employment status (P<0.0001); marital status (P<0.0001); delayed medical care (P = 0.0463); diabetes (P<0.0001); BMI (P = 0.0112); physical activity level (P<0.0001); cholesterol levels (P = 0.009); hypertension (P<0.0001); smoking status (P = 0.303); general health status (P<0.0001); age (P<0.0001); and years of BC survivorship (P<0.0001) in women with BC and CVD compared to women with BC but no CVD.

Table 2. Sociodemographic characteristics of female breast cancer patients.

^**P-value ≤ 0.05.

Logistic regression: modeling CVD risk in women with BC

A forward stepwise logistic regression was used to model the likelihood of a CVD diagnosis against sociodemographic characteristics and health risk behaviors (likelihood ratio = 16.51, P-value < 0.0001; Table 3).

Table 3. Forward stepwise logistic prediction model for CVD among BC survivors.

Forward stepwise prediction model was fit to the BC data (likelihood ratio = 16.51, P-value < 0.0001, c-statistic = 0.754).

^**P-value ≤ 0.05.

^*P-value ≤ 0.1.

Advancing age is clearly a significant risk factor for comorbid CVD among women with BC. Women aged >85 and women between ages 66–84 have an 875% and 324% higher odds of having a CVD compared to those individuals less than 45 years old, respectively (P = 0.0005 95% CI =2.74–34.7; P = 0.0121 95% CI =1.38–13.0). Women who answered “Yes” to a delay in medical care due to cost were shown to have 79% higher odds of reporting a CVD diagnosis (P-value = 0.0073 95% CI = 1.17–2.72) compared to women who answered “No.” Women with hypertension were 106% more likely report a CVD diagnosis (P-value<0.0001 95% CI = 1.50–2,83) compared to individuals with no hypertension. Women who described their general health status as “fair/poor” or “good/very good” have a 705% and 141% increased likelihood of reporting CVD (P<0.0001 95% CI = 4.35–19.9; P = 0.0032 95% CI = 1.35–4.30), respectively compared to those who described their health status as “Excellent.”

Smoking is a well-established risk factor for CVD and cancer. The NHIS survey uses the threshold of 100 cigarettes smoked, in addition to current smoking status to indicate habitual cigarette use. In this analysis 100 cigarettes smoked was an indicator of general smoking history. We note that research specific to whether 100 lifetime cigarettes is a discriminating cut-point for distinguishing smokers from past smokers and never smokers is limited (25). Women who indicated that they had smoked more than 100 cigarettes in their lifetime were 28% more likely to report a CVD diagnosis compared to women who had not smoked more than 100 cigarettes in their lifetime (P = 0.085 95% CI = 0.97–1.70).

Discussion

These results provide additional support for the importance of managing CVD-related risk factors among women with BC. Diabetes, hypertension, and high cholesterol may be managed by easily accessible medications like metformin, propranolol, and atorvastatin. However, preventive behavior support may be necessary to help BC survivors modify lifestyle risk factors such as smoking, poor nutrition, and sedentary behaviors. We also found a direct association between self-reported CVD diagnoses in BC survivorship that rises sharply at 7 years after BC diagnosis and continues to rise beyond 10 years BC survivorship. Future public health considerations should focus on early identification and ongoing management of cardiovascular risks (26).

There was an expected increase in the prevalence of CVD when younger age cohorts were compared to older cohorts. Age is an important independent factor in CVD development in this population, as older patients are more likely to be exposed to more carcinogens over their lifetime and have higher risk of hypertension, diabetes, and hyperlipidemia. However, these findings still underscore the importance of preventive behaviors and long-term cardio-oncology focused care as age and survivorship increases, and CVD risk increases. One of the key concerns undergirding this observational study is the fact that clinical guidelines do not provide enhanced screening or surveillance guidelines for older BC survivors and greater than 5-year BC survivors, beyond that which is status quo for older women without BC.

The associations were confirmed in this study and emphasize the importance of enhanced monitoring of these risk factors in BC survivors across the survivorship history and well-beyond 5 years as recommended by current clinical guidelines. Furthermore, this continues to signal the need for additional research to disentangle the relative importance of individual, clinical, behavioral and population-level risk factors impacting cardiovascular health among long term BC survivors.

Contrary to previous research, insurance status, education, and race/ethnicity were not statistically associated with a CVD diagnosis in this sample of BC survivors. This finding contrasts with previous research showing that Black patients typically have poorer CVD outcomes (6). Racial gaps in BC survival remain a serious concern. Despite being diagnosed less frequently, American Indian, Alaska Native women, and Black women are 38% more likely to die from breast cancer compared to White women. This survival gap appears across disease subtypes and at every stage of progression, except when cancer is localized, where they are already 10% less likely to be diagnosed than White women (58% vs. 68%) (3). However, our findings may be reasonably attributed to a sampling bias. NHIS survey respondents were majority white (84.3% BC survivors without CVD; and 83.2% BC survivors with CVD).

A particularly concerning finding is that BC survivors were 79% more likely to report a CVD if they had delayed medical care for 12 months or more due to cost or other reasons compared to those who had not delayed care. The delay in medical care/not seeing a doctor in more than 12 months raises important implications for adequate survivorship care, risk factor management and recurrent/secondary cancer screening. The decision to delay care highlights how economic barriers at the individual, community and societal levels can have far-reaching health consequences for BC patients (34). Social support and wrap around services may be important for mitigating socio-economic risk factors such as unemployment, underemployment and other factors driving inability or unwillingness to seek routine care. We should continue to prioritize health care policies to promote screenings, prevention and early management of comorbid conditions among BC survivors.

Patient-provider communication is the foundation of CVD education and CVD risk awareness for women with BC. The fact that BC survivors experience delayed care and unmanaged comorbid conditions indicates that BC survivors may not fully understand how their diagnosis and treatment change both acute and long-term cardiovascular risk. Primary care providers (PCPs) who typically see BC patients across the continuum of their care, are key to patient education on the impact of exercise, healthy diet, weight loss, stress management and hypertension control for reducing CVD risk and refer patients to appropriate specialists when appropriate. An exploratory analysis of BC survivors found that while patients desired lifestyle-related information and guidance from their PCPs, there were many times this information was not adequately communicated (27). Early and consistent PCP involvement in care may support better cardio-oncological follow up over time and minimize the likelihood of poor CVD outcomes. Provider communication may empower patients to be proactive in utilizing preventative services and facilitate a greater sense of self-efficacy among patients to address CVD risks (28).

Limitations

This study has several limitations that should be noted. First, the sample lacked racial/ethnic minority diversity as most eligible respondents were white BC survivors. We also note that the overall prevalence of comorbid BC and cardiovascular disease in the larger NHIS sampling pool was relatively low, which may limit generalizability.

Second, reliance on self-reported medical history is subject to recall bias but there is a dearth in research validating the accuracy of self-reported CVD in US populations. Without objective verification through medical records, the data is subject to both underreporting and overreporting depending on the CVD condition. Research in Asian, Swedish, and Australian populations demonstrates high sensitivity and specificity for self-reported cardiovascular events such as stroke and myocardial infarction, though accuracy is lower for conditions like angina and coronary artery disease (29–31, 36).

Third, the cross-sectional study design precludes any determination of temporality or causality, and we are unable to identify timing of specific CVD onset.

Fourth, we also acknowledge that BC stage and treatment type influence subsequent CVD risk but were unable to control for these factors given the data source. Consequently, these findings may be subject to survivorship bias, potentially underrepresenting women with advanced-stage BC who died early in survivorship and those whose BC treatment protocols elevated cardiovascular risk or early mortality.

Furthermore, the cross-sectional study design does not allow an analysis of behavior and/or comorbidities over time and may exclude other factors associated with the quality of life and health outcomes of BC survivors. Future considerations of cultural, structural neighborhood and community-level influences may elucidate determining factors and risks beyond those evaluated in this study.

Implications for practice

There is a clear deficit in consistent implementation of guidelines for long-term follow-up and patient education, despite decades of evidence-based research indicating the need for enhanced, long-term CVD prevention and management (32). Current United States BC guidelines recommend yearly mammograms to BC patients for the first 5 years. Follow up care focuses on having physical exams done every 3–6 months only for the first 3 years and 6–12 months during the fourth and fifth year. After the fifth year, the current guidelines state that BC survivors can return to an annual physical without additional screening (22). While the current guidelines focus on monitoring potential remission of BC, there much less clear guidance on enhanced screening or monitoring for CVD and related. The American Society of Clinical Oncology recommends an echocardiogram within the 6–12-month period after treatment in asymptomatic patients who may be at a higher risk of CVD development (17). Due to the cardiotoxic nature of many of the types of BC treatments as well as the common risk factors between BC and CVD, diseases like hypertension, hyperlipidemia, and diabetes are better managed, if there is synergy between guidelines, clinical care protocols and patient education. CVD screening in breast cancer survivors should be lifelong and risk-stratified based on BC treatment and other health profile indicators. Clinicians should consider prioritizing a holistic approach to prevent CVD in BC survivors by monitoring BC survivors for cardiovascular dysfunction not just immediately after treatment, but for more than a decade afterward. Primary care physicians play a crucial role in coordinating this long-term cardiovascular monitoring, with cardio-oncology or cardiologist support as needed. This comprehensive approach to care can improve long-term survivorship outcomes for BC survivors.

Conclusion

The study also supports previous research regarding the role that socio-demographic and lifestyle factors play in BC survivors' risk of comorbid CVD. This study signals the need for comprehensive, long-term CVD screening for BC survivors by way of stronger CVD related guidelines. The increased risk of CVD beyond 7 years after BC diagnosis necessitates a paradigm shift in survivorship care, emphasizing cardiovascular health alongside cancer surveillance. With adequate screening and risk factor management the CVD-related deaths BC survivors may be significantly reduced.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher's note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Footnotes

Edited by: Iokfai Cheang, Nanjing Medical University, China

Reviewed by: Bayu Begashaw Bekele, Augusta University, United States

Vinit Nalawade, Duke University Health System, United States