A balancing act: racial disparities in cardiovascular disease mortality among women diagnosed with breast cancer

Introduction

Breast cancer remains the most commonly diagnosed malignancy among women and is at the vanguard of precision medicine, with targeted therapies aimed at curing patients of their disease (1). While these therapies are effective at improving breast cancer outcomes, with an average 5-year survival of 90% (2), the cardiotoxic effects of therapies are well-documented (3). Recent evidence suggests that breast cancer survivors are at greater risk for cardiovascular disease (CVD) mortality (4), relative to women without breast cancer, manifesting approximately 7 years after diagnosis (5).

Evidence documenting the long-term effects of these therapies largely comes from clinical trials (6-8). However, breast cancer patients that participate in clinical trials are often healthier than the larger breast cancer population, and without underlying comorbidities (9,10). In addition, minority populations, such as non-Hispanic black (NHB) women, are underrepresented in clinical trial populations (11), and are more likely to present with obesity and other comorbidities at diagnosis compared to their non-Hispanic white (NHW) counterparts (12,13), potentially increasing susceptibility to CVD-related events (14,15).

Information gleaned from population-based observational studies can inform interventions over the course of treatment and follow-up care recommendations for breast cancer patients to mitigate adverse effects of treatment. However, few studies have examined the long-term effects of these therapeutic agents in population-based settings, and those doing so have yielded inconsistent findings (16,17). This may be due to methodologic challenges—such as competing risks and confounding by indication—in observational studies that make the calculation of reliable estimates difficult (3,14). Confounding by indication can occur in studies when treatment is not randomized; women with poorer prognosis are more likely to receive treatments and die of breast cancer, potentially leading to erroneous conclusions that curative treatments are harmful with respect to breast cancer mortality (18). As breast cancer mortality competes with CVD mortality, confounding by indication could potentially explain findings that suggest cardiotoxic treatments are protective against CVD mortality (19). In this study, we used methods to mitigate potential bias due to confounding by indication and competing events to quantify the effect of breast cancer therapies on CVD mortality, and evaluate racial differences in the effect of these therapies on CVD mortality.

Methods

The Georgia Cancer Registry (GCR) is a statewide population-based registry that has collected nearly all cancer cases diagnosed among Georgia residents since January 1, 1995. Using this registry, we identified 3,580 NHB and 4,923 NHW women with a first primary invasive breast cancer diagnosis [2010–2014]. Women were included if they resided in the metropolitan-Atlanta area at the time of diagnosis and excluded if they were <18 years of age or had an autopsy diagnosis. Underlying cause of death was determined directly from death certificates and CVD mortality was defined using the International Classification of Diseases (ICD), tenth revision codes I00-I99 (ICD-9 codes 390-459). The GCR links annually to the State Office of Vital Records to identify in-state deaths, and the US National Death Index to identify deaths that occur outside of Georgia. Follow-up information was available for women through December 31, 2018. This study was conducted in accordance with the Declaration of Helsinki and has been approved by the institutional review board (IRB) of Emory University (IRB00099875) on 24 October 2017. Participant consent was not required due to the registry-based nature of the study.

Patient treatment information is routinely collected by the GCR. For the purposes of this analysis, treatments considered included radiation therapy, chemotherapy, hormone therapy and HER2-targeted therapies (trastuzumab). Individual agents are not captured within the GCR in discrete fields, therefore textual descriptions, required in Georgia for all cancer treatments, were algorithmically searched to identify records that suggested trastuzumab receipt.

We used propensity score weighting to mitigate confounding by indication in analyses of the effect of breast cancer treatments on breast cancer mortality, which allows for more reliable estimation of CVD mortality. This method creates comparable populations balanced on potential confounders, specifically those related to indication for the specified therapies or underlying disease severity (20). Estimates of association for each treatment modality with CVD mortality were calculated using average treatment effect among the treated (ATT)-weights (21). These weights make the covariate distribution among those who did not receive each therapy comparable to those who did receive the therapy. Propensity score models included the following variables for all treatment modalities: age, stage, surgery, radiation therapy, insurance status, poverty level, and race, as well as interaction terms between race and all remaining variables in the model. Models additionally included tumor subtype and chemotherapy in hormone therapy and trastuzumab models, and estrogen receptor (ER) status and lymph node involvement in chemotherapy models. We used standardized differences (22) to assess the covariate balance across treatment within race strata, which were adequate (<0.20) for all treatments except radiation therapy (Tables S1-S4), suggesting that reliable estimates for the association of radiation therapy with CVD mortality could not be computed in our dataset and was excluded from this analysis. ATT-weights were used in Cox proportional hazards models, which censor competing causes of death at the time of death, to estimate the cause-specific hazard ratios (HRs) and 95% confidence intervals (CIs) for the association between each breast cancer treatment modality and CVD mortality, as well as within race group. We also conducted propensity score weighted (based on ATT-weights) analyses using Fine-Gray models, an alternate method that accounts for the presence of competing causes of death by including those who died from competing causes in the risk set (23), to produce subdistribution hazard ratios (sdHR).

Information on comorbidities at diagnosis are not collected in the GCR registry. To account for the potential for unmeasured confounding by comorbidities, namely CVD at diagnosis, we performed a quantitative bias analysis (24,25). To perform the bias analysis, we assigned values to the bias parameters based on existing literature and clinical input for the association between CVD and CVD mortality, the association between CVD and the exposures, and the prevalence of CVD among the source population (i.e., breast cancer survivors). Women with CVD and multiple comorbidities at diagnosis are less likely to receive breast cancer therapies (26), although there is limited literature on the topic, we assigned an HR =0.70 for the bias analysis. A CVD diagnosis and presence of CVD related lifestyle factors are strongly associated with CVD mortality, we assigned an HR =4.2 based on a recent publication (27). Finally, the estimated prevalence of CVD among breast cancer patients is 29.2% (28). Using the above values of bias parameters, we performed a multidimensional bias analysis to account for unmeasured confounding.

In a sensitivity analysis, we repeated the analyses among stage I–III breast tumors, as women diagnosed with stage IV disease are less likely to die from side effects of the treatment, but more likely to die from their disease.

Results

Overall, we observed 172 CVD-related deaths (10% of all deaths) with a similar proportion of CVD-related deaths, out of the total study population at risk, among NHB (2.0%) and NHW women (2.1%) (Table 1). Frequencies of CVD-related deaths according to breast cancer therapy and race/ethnicity can be found in Table S5. NHB women were more likely to receive chemotherapy (57% vs. 40%) and trastuzumab (14% vs. 11%), but less likely to receive hormone therapy (55% vs. 64%) compared to their NHW counterparts. The median follow-up time was 4.7 years (interquartile range =3.4–6.3 years).

In the weighted cohort, the hazard of CVD mortality was similar among women who received chemotherapy (HR =1.10, 95% CI: 0.62, 1.96) and hormone therapy (HR =0.94, 95% CI: 0.59, 1.50), relative to women who did not receive the respective therapies (Table 2). Conversely, the hazard of CVD mortality was somewhat higher among women who received trastuzumab (HR =2.05, 95% CI: 0.76, 5.52).

In models examining racial differences, among women who received chemotherapy, relative to those who did not, we observed a higher hazard of CVD mortality in NHB women (HR =1.45, 95% CI: 0.60, 3.51) but not NHW (HR =0.86, 95% CI: 0.40, 1.88). Similar trends were observed among women who received hormone therapy (NHB: HR =2.18, 95% CI: 0.78, 6.12; NHW: HR =0.66, 95% CI: 0.39, 1.13). Due to a limited number of CVD events among women who received trastuzumab (n=9), we were unable to estimate race-specific associations. Supplemental analyses accounting for the influence of non-CVD deaths on CVD-mortality yielded similar results to primary analyses, except that associations with chemotherapy and hormone therapy with CVD mortality were slightly stronger among NHB women (Table S6).

In our quantitative bias analysis accounting for the possible unmeasured confounding due to comorbid conditions such as CVD at diagnosis, results suggest that there was a slight bias in the negative direction and that adjusting for the presence of CVD at diagnosis would strengthen the observed estimates of association (Table 3). Additionally, results were similar after excluding women diagnosed with stage IV breast cancer, though less precise (Table S7).

Discussion

In this population-based study, we observed a higher hazard of CVD mortality among women who received trastuzumab. We also observed higher hazards of CVD mortality among NHB women who received chemotherapy or hormone therapy, but not among NHW women. Our findings may have important clinical implications, as the results suggest that clinicians should consider the best strategy to provide curative treatment for breast cancer patients, while simultaneously minimizing potential treatment-related cardiotoxicities, particularly among NHB women.

To our knowledge, our study is the first population-based study to report higher hazards of CVD mortality among women who received trastuzumab, compared to women who did not. Our findings could be due to chance, as the prevalence of trastuzumab receipt was low and few CVD deaths occurred over the course of study follow-up. Previously, an observational study conducted among European women diagnosed with HER2-positive early breast cancer and treated with trastuzumab found most cardiac events were asymptomatic or mild and few women died due to cardiac-related events (0.2%) (29). In contrast, we observed a higher incidence of CVD mortality (~2%), likely due to differences in study populations, as the prevalence of heart disease in the US is almost twice the prevalence in Europe (30). Evidence from clinical trials suggest trastuzumab is associated with left ventricular dysfunction and heart failure, although the overall incidence of events appears low (0% to 7.2%) and mostly reversible (6,31-33). However, participants in clinical trials often differ from women in a real-world setting with respect to age, race, and comorbid conditions (9,10), all of which are risk factors for CVD-related events that could increase susceptibility to the cardiotoxic effects of treatment. Previous population-based studies conducted in the US have suggested higher rates of trastuzumab-related cardiotoxicity than those reported in clinical trials, particularly among older breast cancer survivors and those with underlying conditions (34,35). The well-recognized toxic effects of trastuzumab have led clinicians and researchers to reconsider the optimal duration of treatment, and recent evidence suggests that among women with HER2-positive early breast cancer, shortening the treatment duration from 12 to 6 months resulted in similar efficacy, while reducing cardiotoxicities and other adverse events (36).

In this study, we observed higher hazards of CVD mortality among NHB women who received chemotherapy or hormone therapy, but not among NHW women. In our previous analysis, we reported higher hazards of CVD mortality among NHB women relative to their NHW counterparts (15). The present findings suggest that treatment may contribute to the previously observed disparities in CVD mortality among breast cancer survivors. NHB women are more likely to be diagnosed with aggressive tumors (37) and with a higher stage disease (38). As a result, NHB women are more likely to receive anthracycline-based chemotherapy regimens, which may explain the disparate results, as anthracyclines have known cardiotoxic effects (3,39). Evidence from clinical trials suggest that hormone therapy use, particularly long-term use of aromatase inhibitors (AIs), may increase risk of CVD-related events, though the overall incidence of events remains low (3,40). However, in one of the trials, the risk of CVD-related events among women treated with AIs was higher among women with pre-existing heart disease (17% in the anastrozole arm compared to 10% in the tamoxifen arm) (41), resulting in a recommendation from the US Food and Drug Administration to weigh both the risks and benefits of anastrozole use in this subset of patients (42). The observed association among NHB women who received hormone therapy in our study may be due, in part, to the presence of underlying comorbidities (e.g., obesity and hypertension) that are more common among NHB women, especially those with hormone receptor positive disease. Other common risk factors for both CVD and breast cancer include age, hormone replacements, diet, tobacco use, alcohol intake, and physical activity (3). However, in our quantitative bias analysis assessment, the bias due to unmeasured confounding by underlying comorbidities likely biased the results in the negative direction and would not account for our findings.

This prospective study is the first population-based study to examine differences in various treatment-associated CVD mortality among NHB and NHW women. This is particularly important as women with preexisting comorbidities, including CVD-related risk factors that increase the risk of treatment-associated cardiotoxicity and are more common among NHB women, are often excluded from clinical trials. This study is limited by the breast cancer therapies ascertained from the GCR, which reports under ascertainment of treatments that may have led to misclassification of the exposure, and by lack of information on many known cardiovascular risk factors. In addition, we were unable to determine the specific chemotherapeutic agents or the cumulative dosage, important factors in determining chemotherapy-related cardiotoxicity. Lastly, estimates of association reported in this study for all treatments are imprecise due to the methodological limitations of the approach and the limited number of outcomes in the relatively short follow-up.

Our results suggest that receipt of trastuzumab may increase the risk of CVD mortality among women with breast cancer and that chemotherapy and hormone therapy may increase the risk of CVD mortality among NHB women. Among US breast cancer survivors, rates of mortality due to CVD-related events are higher than that of breast cancer, particularly among NHB women (3), highlighting the importance of managing both CVD and cancer. Breast cancer patients, especially NHB women, would benefit from clinician assessment of cardiac risk profiles to identify and manage CVD risk factors and guide treatment decisions that balance the need for both curative cancer treatment and prevention of CVD-related events and mortality.

Acknowledgments

Funding: The collection of cancer incidence data used in this study was supported by contract HHSN261201800003I, Task Order HHSN26100001 from the NCI and cooperative agreement 5NU58DP003875-04 from the CDC. Lindsay J Collin was supported, in part, by the US National Cancer Institute (F31CA239566). Funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Footnote

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at http://dx.doi.org/10.21037/ace.2020.01.02). KG reports Emory receives institutional funding to support my role as principal investigator on clinical trials sponsored by industry; including Merck, Genentech, Pfizer, and Seattle Genetics. KW reports SEER Contract from NCI, during the conduct of the study. The other authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. This study was conducted in accordance with the Declaration of Helsinki (as revised in 2013) and has been approved by the institutional review board (IRB) of Emory University (IRB00099875) on 24 October 2017. Participant consent was not required due to the registry-based nature of the study.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. DeSantis CE, Ma J, Goding Sauer A, et al. Breast cancer statistics, 2017, racial disparity in mortality by state. CA Cancer J Clin 2017;67:439-48. [Crossref] [PubMed]
2. Breast Cancer Facts & Figures 2019-2020. Atlanta: American Cancer Society, Inc 2019.
3. Mehta LS, Watson KE, Barac A, et al. Cardiovascular Disease and Breast Cancer: Where These Entities Intersect: A Scientific Statement From the American Heart Association. Circulation 2018;137:e30-e66. [Crossref] [PubMed]
4. Coughlin SS, Kapuku G. Cancer Survivorship and Subclinical Myocardial Damage. Am J Epidemiol 2019; [Epub ahead of print]. [Crossref] [PubMed]
5. Bradshaw PT, Stevens J, Khankari N, et al. Cardiovascular Disease Mortality Among Breast Cancer Survivors. Epidemiology 2016;27:6-13. [Crossref] [PubMed]
6. Cameron D, Piccart-Gebhart MJ, Gelber RD, et al. 11 years' follow-up of trastuzumab after adjuvant chemotherapy in HER2-positive early breast cancer: final analysis of the HERceptin Adjuvant (HERA) trial. Lancet (London, England) 2017;389:1195-205. [Crossref] [PubMed]
7. Goldvaser H, Barnes TA, Šeruga B, et al. Toxicity of Extended Adjuvant Therapy With Aromatase Inhibitors in Early Breast Cancer: A Systematic Review and Meta-analysis. J Natl Cancer Inst 2018; [Crossref] [PubMed]
8. Long HD, Lin YE, Zhang JJ, et al. Risk of Congestive Heart Failure in Early Breast Cancer Patients Undergoing Adjuvant Treatment With Trastuzumab: A Meta-Analysis. Oncologist 2016;21:547-54. [Crossref] [PubMed]
9. Elting LS, Cooksley C, Bekele BN, et al. Generalizability of cancer clinical trial results: prognostic differences between participants and nonparticipants. Cancer 2006;106:2452-8. [Crossref] [PubMed]
10. Murthy VH, Krumholz HM, Gross CP. Participation in cancer clinical trials: race-, sex-, and age-based disparities. JAMA 2004;291:2720-6. [Crossref] [PubMed]
11. Ford JG, Howerton MW, Lai GY, et al. Barriers to recruiting underrepresented populations to cancer clinical trials: a systematic review. Cancer 2008;112:228-42. [Crossref] [PubMed]
12. Jemal A, Robbins AS, Lin CC, et al. Factors That Contributed to Black-White Disparities in Survival Among Nonelderly Women With Breast Cancer Between 2004 and 2013. J Clin Oncol 2018;36:14-24. [Crossref] [PubMed]
13. Tammemagi CM, Nerenz D, Neslund-Dudas C, et al. Comorbidity and survival disparities among black and white patients with breast cancer. JAMA 2005;294:1765-72. [Crossref] [PubMed]
14. Gulati M, Mulvagh SL. The connection between the breast and heart in a woman: Breast cancer and cardiovascular disease. Clin Cardiol 2018;41:253-7. [Crossref] [PubMed]
15. Troeschel AN, Liu Y, Collin LJ, et al. Race differences in cardiovascular disease and breast cancer mortality among US women diagnosed with invasive breast cancer. Int J Epidemiol 2019;48:1897-905. [Crossref] [PubMed]
16. Bowles EJA, Wellman R, Feigelson HS, et al. Risk of heart failure in breast cancer patients after anthracycline and trastuzumab treatment: a retrospective cohort study. J Natl Cancer Inst 2012;104:1293-305. [Crossref] [PubMed]
17. Gernaat SAM, Ho PJ, Rijnberg N, et al. Risk of death from cardiovascular disease following breast cancer: a systematic review. Breast Cancer Res Treat 2017;164:537-55. [Crossref] [PubMed]
18. Bosco JLF, Silliman RA, Thwin SS, et al. A most stubborn bias: No adjustment method fully resolves confounding by indication in observational studies. J Clin Epidemiol 2010;63:64-74. [Crossref] [PubMed]
19. Bouchardy C, Rapiti E, Usel M, et al. Excess of cardiovascular mortality among node-negative breast cancer patients irradiated for inner-quadrant tumors. Ann Oncol 2010;21:459-65. [Crossref] [PubMed]
20. Brookhart MA, Wyss R, Layton JB, et al. Propensity score methods for confounding control in nonexperimental research. Circ Cardiovasc Qual Outcomes 2013;6:604-11. [Crossref] [PubMed]
21. Austin PC. The use of propensity score methods with survival or time-to-event outcomes: reporting measures of effect similar to those used in randomized experiments. Stat Med 2014;33:1242-58. [Crossref] [PubMed]
22. Austin PC. The relative ability of different propensity score methods to balance measured covariates between treated and untreated subjects in observational studies. Med Decis Making 2009;29:661-77. [Crossref] [PubMed]
23. Fine JP, Gray RJ. A proportional hazards model for the subdistribution of a competing risk. J Am Stat Assoc 1999;94:496-509. [Crossref]
24. Lash TL, Fox MP, MacLehose RF, et al. Good practices for quantitative bias analysis. Int J Epidemiol 2014;43:1969-85. [Crossref] [PubMed]
25. Flanders WD, Khoury MJ. Indirect assessment of confounding: graphic description and limits on effect of adjusting for covariates. Epidemiology 1990;1:239-46. [Crossref] [PubMed]
26. Kimmick G, Fleming ST, Sabatino SA, et al. Comorbidity burden and guideline-concordant care for breast cancer. J Am Geriatr Soc 2014;62:482-8. [Crossref] [PubMed]
27. Yang Q, Cogswell ME, Flanders WD, et al. Trends in cardiovascular health metrics and associations with all-cause and CVD mortality among US adults. JAMA 2012;307:1273-83. [Crossref] [PubMed]
28. Patnaik JL, Byers T, Diguiseppi C, et al. The influence of comorbidities on overall survival among older women diagnosed with breast cancer. J Natl Cancer Inst 2011;103:1101-11. [Crossref] [PubMed]
29. Lidbrink E, Chmielowska E, Otremba B, et al. A real-world study of cardiac events in > 3700 patients with HER2-positive early breast cancer treated with trastuzumab: final analysis of the OHERA study. Breast Cancer Res Treat 2019;174:187-96. [Crossref] [PubMed]
30. Thorpe KE, Howard DH, Galactionova K. Differences in disease prevalence as a source of the U.S.-European health care spending gap. Health Aff (Millwood) 2007;26:w678-86. [Crossref] [PubMed]
31. Joensuu H, Bono P, Kataja V, et al. Fluorouracil, epirubicin, and cyclophosphamide with either docetaxel or vinorelbine, with or without trastuzumab, as adjuvant treatments of breast cancer: final results of the FinHer Trial. J Clin Oncol 2009;27:5685-92. [Crossref] [PubMed]
32. Perez EA, Romond EH, Suman VJ, et al. Four-year follow-up of trastuzumab plus adjuvant chemotherapy for operable human epidermal growth factor receptor 2-positive breast cancer: joint analysis of data from NCCTG N9831 and NSABP B-31. J Clin Oncol 2011;29:3366-73. [Crossref] [PubMed]
33. Slamon D, Eiermann W, Robert N, et al. Adjuvant trastuzumab in HER2-positive breast cancer. N Engl J Med 2011;365:1273-83. [Crossref] [PubMed]
34. Chavez-MacGregor M, Zhang N, Buchholz TA, et al. Trastuzumab-related cardiotoxicity among older patients with breast cancer. J Clin Oncol 2013;31:4222-8. [Crossref] [PubMed]
35. Chen J, Long JB, Hurria A, et al. Incidence of heart failure or cardiomyopathy after adjuvant trastuzumab therapy for breast cancer. Journal of the American College of Cardiology 2012;60:2504-12. [Crossref] [PubMed]
36. Earl HM, Hiller L, Vallier AL, et al. 6 versus 12 months of adjuvant trastuzumab for HER2-positive early breast cancer (PERSEPHONE): 4-year disease-free survival results of a randomised phase 3 non-inferiority trial. Lancet (London, England) 2019;393:2599-612. [Crossref] [PubMed]
37. Agurs-Collins T, Dunn BK, Browne D, et al. Epidemiology of health disparities in relation to the biology of estrogen receptor-negative breast cancer. Seminars in Oncology 2010;37:384-401. [Crossref] [PubMed]
38. Newman LA. Breast cancer in African-American women. The Oncologist 2005;10:1-14. [Crossref] [PubMed]
39. Smith LA, Cornelius VR, Plummer CJ, et al. Cardiotoxicity of anthracycline agents for the treatment of cancer: systematic review and meta-analysis of randomised controlled trials. BMC Cancer 2010;10:337. [Crossref] [PubMed]
40. Amir E, Seruga B, Niraula S, et al. Toxicity of adjuvant endocrine therapy in postmenopausal breast cancer patients: a systematic review and meta-analysis. J Natl Cancer Inst 2011;103:1299-309. [Crossref] [PubMed]
41. Tannock IF. 10-year analysis of the ATAC trial: wrong conclusion? Lancet Oncol 2011;12:216-7. [Crossref] [PubMed]
42. US Food and Drug Administration Highlights of prescribing information. 2009.