Risk of death from cardiovascular disease following breast cancer: a systematic review

Purpose Breast cancer incidence and survival is high, which results in high prevalence of breast cancer survivors. The risk of (death from) cardiovascular disease (CVD) is higher in patients exposed to cardiotoxic treatments, in particular if they have pre-existing CVD risk factors. This study systematically summarized the risk of death from CVD following breast cancer. Methods Databases of Medline, Embase, and the Cochrane Library were systematically searched using the following terms and synonyms: breast cancer, cardiovascular disease, and cause of death. Articles reporting on both risk and risk factors of CVD mortality following breast cancer were eligible for inclusion. The methodological quality of each article was assessed using the Newcastle Ottawa quality assessment scale for cohort studies. Results Fourteen articles were included assessing the risk of CVD mortality among 1,217,910 women with breast cancer. The methodological quality was high for the majority of the studies. Studies were heterogeneous in design, study population, length of follow-up, CVD outcomes, and risk factors. 1.6–10.4% of all women with breast cancer died of CVD. Women with breast cancer had a higher risk of CVD mortality than women from the general population. The risk of CVD mortality was higher among women with breast cancer with older age at diagnosis, left-sided tumor, diagnosis in an earlier calendar period, and black ethnic origin. Conclusions CVD is an important cause of death following breast cancer. Identification of patients at high risk of CVD is important to optimize CVD prevention and tailor breast cancer treatment.


Introduction
Breast cancer incidence has increased substantially over the last decades [1,2], which, in combination with improved survival rates attributable to the availability of screening methods and effective treatments of early and more advance breast cancer [3,4], leads to an increasing number of breast cancer survivors. Cardiovascular disease (CVD) is an important cause of death among these women as the risk of CVD may be increased by cardiotoxic treatments and CVD risk factors [5][6][7][8].
The risk of (death from) CVD following breast cancer is increased in women exposed to cardiotoxic treatments such as mediastinal and left-sided radiotherapy, anthracyclinebased chemotherapy, and trastuzumab, and is even higher in patients with pre-existing CVD risk factors such as diabetes and hypertension [9][10][11][12]. With the current high breast cancer survival rates, especially for women with lower stages, and the large number of women with breast cancer receiving intensive treatment regimens, it is increasingly important to identify patients at high risk of CVD, and to balance the benefits of breast cancer treatment for achieving tumor control with the risks of cardiac toxicity inducing CVD.
As an overview of the available evidence on the risk of dying of CVD in women with breast cancer is currently lacking, we systematically reviewed the literature on the risk and risk factors of death from CVD following breast cancer.

Methods and materials
This systematic review was conducted in accordance with the Preferred Reporting Item for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [13].
A systematic search was performed, and last updated on April 1, 2017, to identify all studies reporting on the risk and risk factors of death from CVD following breast cancer. Databases of Medline (via PubMed), Embase, and the Cochrane Library were systematically searched using the following terms and their synonyms in the search strategy: breast cancer, cardiovascular disease, and cause of death (Table 1). No limits were used.
Articles reporting on both risk and risk factors of CVD mortality in breast cancer patients were eligible for inclusion. Articles with the following criteria were excluded: (1) published before 1990, (2) written in another language than English or Dutch, (3) case reports, reviews, or abstracts. Cross-referencing was performed.

Selection of studies and data extraction
After removal of duplicates, all titles and abstracts of the remained retrieved articles were screened. Abstracts that seemed potentially relevant, based on the in-and exclusion criteria, were screened for full text. The full text of these articles were assessed for eligibility by three investigators independently (S.A.M. Gernaat, P.J. Ho, and N. Rijnberg). Data were extracted using standardized data extraction forms and any disagreements were resolved by discussion. We extracted data on study size, characteristics of breast cancer patients (age, ethnic origin, year of diagnosis, years of follow-up), study design, International Classification of Diseases (ICD) codes for CVD mortality, absolute risk of death from CVD, absolute risk of death from breast cancer, absolute risk of death from any cause, statistical methods used to assess which factors increase the risk of death from CVD, and the risk of CVD mortality per risk factor.

Quality assessment
The methodological quality for each article was assessed by two authors independently (S.A.M. Gernaat and P.J. Ho) using the Newcastle Ottawa Quality Assessment Scale (NOS) for cohort studies [14]. The NOS consists of six multiple-choice questions that address subject selection, comparability, and the assessment of the outcome (i.e. CVD mortality), which sum up to a maximum score of seven. In the present study, a high score on one of these sections indicated that the maximum score (i.e. two for selection and comparability, and three for outcome) was achieved. In all other cases, the study received a low score on that particular section.

Results
The systematic search yielded 10,170 citations including 5911 unique articles, which were screened for title and abstract using the predefined inclusion and exclusion criteria (Fig. 1). After screening the full text of 39 articles, 27 were excluded for the following reasons: published before the calendar year 1990 (n = 3) or articles that did not report the risk and risk factors of death from CVD (n = 24). Crossreferencing identified two additional papers. In total, 14 articles were included in the current systematic review, including 4,773,576 women of which 1,217,910 were diagnosed with breast cancer [5,6,8,[15][16][17][18][19][20][21][22][23][24][25].

Quality assessment
The majority of the studies had the maximum score on the quality assessment for selection, comparability, and outcome [5,8,17,18,[20][21][22]26] (Fig. 2). The study by Nichols et al. [15] had a low score on selection as the study population was a selected group of in situ or invasive breast cancer patients and breast cancer ascertainment was by written self-report. The studies by Berkman et al. [6], Darby et al. [23], and Giordano et al. [24] had low scores on comparability as the hazard ratios (HRs) were not adjusted for factors other than age at diagnosis, the CVD mortality rates were unadjusted, and the HRs were only adjusted for other factors than age at diagnosis, respectively. The studies by Hooning et al. [25] and McCullough et al. [16] had low scores on the outcome attainment as the assessment of CVD deaths was by hospital records and subjects were lost to follow-up or the follow-up rate was less than 70%, respectively.

Cardiovascular disease mortality in breast cancer patients compared with the general population
Bradshaw et al. [8] included 1413 women with primary in situ or invasive breast cancer diagnosed in the United States (U.S.) between 1996 and 1997, and 1411 age-matched women from the general population (Table 2). Mean age at breast cancer diagnosis and reference date for women from the general population were 59 and 57 years, respectively. During the follow-up time, which ranged between 0.2 and 13.5 years for both groups, 9.4% of women with breast cancer and 7.4% of women from the general population died of CVD. After adjusting for age, menopausal status, and other CVD risk factors, women with breast cancer had a 1.9 (95% confidence interval (CI) = 1.4-2.7) times higher risk to die of CVD after at least 7 years post diagnosis than women from the general population (Table 3). Riihimäki et al. [5] included all 3,676,472 female Swedish residents born before 1977 (Table 2). Of these, 122,217 were diagnosed with primary invasive breast cancer between 1987 and 2006. During a maximum followup of 19 years, 10.4% and 7.5% of women died of CVD, respectively. Women with breast cancer had a 1.14 (95% CI 1.10-1. 19) times higher risk to die of coronary heart disease, a 1.29 (95% CI 1.22-1.37) times higher risk to die of heart failure, and a 1.24 (95% CI 1.17-1.32) times higher risk to die of other heart disease than women from the general population, independent of age, socioeconomic index, and geographical region of residence in Sweden (Table 3).
Cardiovascular disease mortality in breast cancer patients by patient, tumor, and treatment characteristics Colzani et al. [18] included 12,850 Swedish women younger than 75 years of age at diagnosis with primary invasive stage I to III breast cancer between 1990 and 2006 ( Table 2). During a maximum follow-up of ten years, 1.8% of all women died of CVD. After adjusting for clinical, tumor, and treatment characteristics, except the one of interest, women with breast cancer were at increased risk of CVD mortality if they were older at diagnosis (65-74 years vs. 45-54 years: hazard ratio (HR) = 17.9, 95% CI 8.0-39.7), if diagnosed in an earlier calendar period (1990-1994 vs. 2000-2006: HR = 2.1, 95% CI 1.2-3.6), and treated with only surgery (HR = 2.1, 95% CI 1. 2-3.8) or surgery in combination with hormonal therapy (HR = 2.2, 95% CI 1.5-3.2) compared with surgery in combination with radiotherapy and hormonal therapy ( Table 3).
Hooning et al. [25] included 7425 women younger than 71 years of age at diagnosis with primary invasive stage I to IIIA breast cancer in the Netherlands between 1970 and 1986 ( Table 2). During a median follow-up of 13.8 years, 5.3% of all women died of CVD. After adjusting for clinical, tumor, and treatment characteristics, women with breast cancer were at increased risk of CVD mortality with each year increase in age at diagnosis (HR = 1.12, 95% CI 1.10-1.14), if diagnosed in an earlier calendar period (1976-1980 vs. 1981-1986: HR = 1.54, 95% CI 1.11-2.14), and treated with a combination of surgery and radiotherapy compared with only surgery (HR = 2.03, 95% CI 1.33-3.10) ( Table 3).

Cardiovascular disease mortality in breast cancer patients by laterality of the tumor
Bouchardy et al. [20] included 1245 women with a mean age of 57.4 years at diagnosis with primary lymph nodenegative breast cancer in Switzerland between 1980 and 2004 ( Table 2). During a mean follow-up of 7.7 years, 2.2% of all women died of CVD. Among women treated with radiotherapy, an inner quadrant tumor was associated with a 2.46 (95% CI 1.13-5.37) higher risk of dying of CVD, adjusted for clinical, tumor, and treatment characteristics (Table 3).  2 Quality assessment by the Newcastle Ottawa Quality Assessment Scale Selection was based on the representativeness of the breast cancer cohort and ascertainment of breast cancer. Comparability was based on the comparability of cohorts on the basis of the design or analysis. Outcome was based on the assessment of death from cardiovascular disease, on the length of follow-up (C10 years), and adequacy of follow-up of the cohorts. A high score on one of these sections indicated that the maximum score on that particular section (i.e. two for selection and comparability, and three for outcome) was achieved. In all other cases, the study received a low score on that particular section Darby et al. [23] included 308,861 women between 20 and 79 years of age at diagnosis with primary in situ or invasive breast cancer in the USA between 1973 and 2001 ( Table 2). During a maximum follow-up of 29 years, 4.2% of all women died of CVD. In women treated with radiotherapy and diagnosed between 1973 and 1982, left-sided breast cancer led to higher mortality ratios (MR) ten to 14 years post diagnosis (unadjusted MR = 1.42, 95% CI 1.11-1.82) and over 15 years post diagnosis (unadjusted MR = 1.58, 95% CI 1.29-1.95) compared with right-sided breast cancer (Table 3). More than ten years post diagnosis, women with left-sided breast cancer had a higher risk of death from CVD (unadjusted MR = 1.44, 95% CI 1.26-1.65), acute myocardial infarction (unadjusted MR = 1.43, 95% CI 1.10-1.87), and other ischemic CVD (unadjusted MR = 1.60, 95% CI 1.26-2.02) compared with women with right-sided breast cancer.
Giordano et al. [24] included 24,785 women with primary in situ or invasive breast cancer diagnosed in the US between 1973 and 1988 ( Table 2). Mean age at diagnosis was 56.9 years (standard deviation (SD) = 13.2) at diagnosis. Eight years post diagnosis, women with left-sided breast cancer who were diagnosed in 1979 had a (unadjusted) 1.50 (95% CI 1.15-1.87) times higher risk to die of CVD compared with women with right-sided breast cancer diagnosed in the same year (Table 3).
Haque et al. [26] included 140,914 women of all ages with ductal carcinoma in situ (DCIS) in the US between 1973 and 2002 ( Table 2).The median follow-up was 11.5 years (interquartile range = 6.8-15.1). Among women diagnosed between 1973 and 1982, a left-sided tumor was associated with a 1.30 (95% CI 1.18-1.42) higher risk of dying of CVD than a right-sided tumor, independent of clinical, tumor, and treatment characteristics. This association was not found for women diagnosed in a more recent calendar period.
Merzenich et al. [21] included 11,982 women with a mean age of 59 years (range = 18-101) at primary diagnosis of in situ or invasive breast cancer in Germany between 1998 and 2008 (Table 2) [21]. During a median               follow-up of 6.5 years (range = 0-15), 2.3% of all women died of CVD. Women with left-sided breast cancer did not have a higher risk of dying of CVD than women with rightsided breast cancer, irrespectively of radiotherapy treatment (Table 3). Among women treated with radiotherapy, women with a history of CVD had a 1.73 times (95% CI 1.11-2.68) higher risk of dying of CVD than women without a history of CVD.

CVD mortality in breast cancer patient by ethnic origin
Berkman et al. [6] included 54,518 white and 6113 black women over 40 years of age at diagnosis with primary DCIS in the US between 1978 and 2010 (  Table 2). Median age at breast cancer diagnosis was 61.2 years (SD = 13.7) for non-Hispanic white women and 56.3 years (SD = 13.1) for Asian and Pacific Islander women. The median follow-up for non-Hispanic white women was 4 years (range = 2-6), during which 5.5% of women died of CVD. The median follow-up for Asian and Pacific Islander women was 3 years (range 2-5), during which 2.6% of women died of CVD. After adjusting for patient, tumor, and registry characteristics, Asian and Pacific Islander women with breast cancer had a HR of 0.77 (95% CI 0.71-0.83) for death from CVD compared to non-Hispanic white women with breast cancer (Table 2). Furthermore, US born Asian and Pacific Islander women with breast cancer had a 1.29 (95% CI 1.08-1.54) times higher risk of death from CVD compared to non-US born Asian and Pacific Islander women with breast cancer.

CVD mortality in breast cancer patients by diet, body weight, and health-behaviors
McCullough et al. [16] included 4452 women diagnosed with primary invasive breast cancer in Switzerland between 1992 and 2011 who had scored their diet according to the American Cancer Society (AMC) guidelines before breast cancer diagnosis, and of these, 2152 women scored their diet also at least one year after breast cancer diagnosis ( Table 2). The AMC guidelines recommend following the general food-based guidelines for primary cancer prevention, which includes eating a plantbased diet rich in vegetables and fruits, whole grains, and which is limited in red and processed meats [27]. Mean age at diagnosis was 70.7 years (SD = 7.2). During a mean follow-up of 9.8 years (SD = 4.9), 5.2% of all women died of CVD. After adjusting for tumor, treatment, and patient characteristics, both pre-diagnostic and post-diagnostic higher AMC diet scores, indicating an unhealthier diet, were not associated with a higher risk of CVD mortality following breast cancer compared with the lowest diet score category (0-2), indicating a healthier diet (Table 3).
Nichols et al. [15] included 5791 women with primary in situ or invasive breast cancer diagnosed in the US between 1988 and 1999 ( Table 2). Mean age at diagnosis was 58.4 years (SD = 10.0). During a mean follow-up of 6.4 years (SD = 1.2), 1.6% of all women died of CVD. After correcting for age, menopausal status, and CVD risk factors, Nichols et al. found a 4.15 (95% CI 1.44-12.0) and 2.45 (95% CI 1.46-4.11) times higher risk of death from CVD in women with a pre-diagnosis underweight (body mass index (BMI; kg/m 2 ): \18.5) and obesity (BMI: C 30), respectively, compared with women with a pre-diagnosis normal weight (BMI: 18.5-24.9) ( Table 3).
Veal et al. [22] included 1925 women aged between 20 and 74 years at diagnosis of DCIS in the US between 1997 and 2006 (Table 2). During a mean follow-up of 6.7 years, 1.8% of all women died of CVD. More hours per week of physical activity before the breast cancer diagnosis was associated with 0.83 (95% CI 0.70-0.98) lower risk of dying of CVD (Table 3).

Discussion
In this review, we systematically summarized the evidence on the risk and risk factors of death from CVD following breast cancer.
The absolute risk of dying of CVD following breast cancer ranges from 1.6% [15] to 10.4% [5], and the risk of CVD mortality is higher in women with breast cancer than women from the general population [5,8]. Older age at diagnosis [6,18,19,25], left-sided tumor [23,24,26], diagnosis in an earlier calendar period [18,25], and black ethnic origin [6] are risk factors of CVD mortality following breast cancer.
Several mechanisms are proposed for the increased risk of CVD mortality in women with breast cancer. CVD risk factors, such as obesity and diabetes, may be more present among breast cancer survivors than women from the general population as breast cancer and CVD have shared risk factors [28]. Also, cardiotoxic effects of breast cancer treatments, specifically mediastinal and left-sided radiotherapy, anthracycline-based chemotherapy, and trastuzumab, are well documented to increase the risk of CVD [29][30][31][32].
In the current review, studies with longer follow-up, i.e. over 10 years [5,6,8], reported higher absolute risks of CVD mortality. The risk of CVD increases with time since diagnosis probably due to increasing age and cardiotoxicity of breast cancer treatments that become apparent after several years [29]. Age is a well-known risk factor for CVD [33][34][35][36], and therefore, expected to be found as a risk factor in women with breast cancer [6,18,19,25]. Schonberg et al. [14] found that 26-40% of older women diagnosed with early-stage breast cancer died of CVD, indicating that the risk of CVD is high in specific subgroups and particularly in older women.
The association between left-sided breast cancer and radiotherapy treatment with a higher risk of CVD mortality was found among women diagnosed in the early 1980's [23,24,26]. Radiotherapy treatment was more cardiotoxic in these years as it usually involved higher doses with large irradiation fields irradiating parts of the heart [37,38]. This may also explain the increased risk of CVD mortality among breast cancer patients diagnosed in an earlier calendar period [18,25]. Colzani et al. [18] did not find an increased risk of CVD mortality among women treated with radiotherapy and/or chemotherapy. Although the baseline risk of CVD was not reported, this result is probably due to patient selection, i.e. women who did not undergo radiotherapy and/or chemotherapy probably had a higher risk of CVD at baseline. The lower risk of CVD in Asian populations [17] and the higher risk of CVD in black populations [6] are reported by several studies and can be explained by the lower and higher presence of CVD risk factors, respectively, such as high blood pressure, obesity, and lipid levels [39][40][41][42].
The present systematic review shows that there are only a limited number of studies investigating the risk and risk factors of CVD mortality following breast cancer, and that these studies are heterogeneous in design, study population, and length of follow-up. Also, the determinants and outcomes, in terms of CVD risk factors and death due to CVD, respectively, vary. We acknowledge that, due to the heterogeneous designs of the included studies, we were unable to perform a meta-analysis, which limited the strength of evidence. Besides limitations, the current review has strengths. This is the first study that systematically summarized the literature on the risk and risk factors of death from CVD following breast cancer. Furthermore, the current systematic review includes a large variety of risk factors of death from CVD in women with breast cancer.
To conclude, the combination of high breast cancer incidence, improved breast cancer survival, presence of CVD risk factors, and cardiotoxic breast cancer treatments has resulted into many breast cancer survivors at risk of CVD. Therefore, it is important to understand the incidence and etiology of CVD in these survivors. Furthermore, identification of women with breast cancer at high risk of CVD is important to minimize the number of women suffering and/or dying of CVD after breast cancer treatment and improve quality of life and long-term prognosis. Clinicians should be able to identify breast cancer patients at increased risk of CVD and provide accurate recommendations for CVD risk reduction strategies specifically for breast cancer survivors at high risk of CVD. The current systematic review, in combination with a recent guideline by Armenian et al. [29] on the prevention and monitoring of cardiac dysfunction in survivors of adult cancers, may help clinicians with such a recommendation. In addition, there are studies investigating the identification of women with breast cancer at high risk of CVD using other measurements, for example, by measuring the coronary artery calcification on radiotherapy planning computed tomography scans [43]. This may further help clinicians with identification of breast cancer patients at high risk of CVD. Identification of breast cancer patients at high risk of CVD is important to optimize CVD prevention of (irreversible) cardiac damage, by adjusting breast cancer treatments accordingly and initializing CVD (preventative) treatment. Furthermore, a tailored individual approach with early and late monitoring of cardiac dysfunction in breast cancer survivors should be implemented in routine care [44].
Funding The current project was conducted within the framework 'Strategic PhD Partnership Program' from the Board of Directors of the University Medical Center Utrecht.

Compliance with ethical standards
Conflict of interest The authors declare that they have no conflict of interest.
Ethical approval For this type of study format, informed consent is not required. This article does not contain any studies with human participants or animals performed by any of the authors.
Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://crea tivecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.