Introduction

With earlier detection and improved treatments, 86 % of New Zealand women diagnosed with breast cancer will survive more than 5 years [1]. Similarly, survival rates now exceed 90 % in Europe [2, 3] and the United States [4]. Among these women, mortality due to cardiovascular disease (CVD) is now more common than breast cancer-related death [5]. In a recent study, 45 % of all CVD-related deaths in breast cancer survivors >55 years of age were attributed to coronary artery disease (CAD), with 20 % the result of cerebrovascular events [5]. In particular, CVD mortality is higher in older breast cancer survivors and those with lymph node involvement [5]. Peak mortality due to circulatory system disorders appears at approximately 5 years post diagnosis [5]; however, higher incidence of cardiac dysfunction has been observed 10–15 years post treatment completion in survivors who have had anthracycline-based chemotherapy and/or radiation therapy [6, 7].

Currently, cardiovascular function is monitored frequently in women who are undergoing trastuzumab (Herceptin) treatment and in those with high risk for CVD while on anthracycline therapies. However, cardiovascular-related dysfunction may not appear until 10–15 years later, suggesting that ongoing monitoring is warranted [7, 8]. Given that the majority of breast cancer diagnoses are in postmenopausal women [9], many of whom may have pre-existing CVD-related risk factors or comorbidity [10], routine clinical cardiovascular health assessment guidelines should be developed.

Post-therapy complications

A number of cancer treatments are associated with increased risk for cardiovascular complications [11–15]. Here we highlight those therapies with known cardiac toxicity that are most frequently administered [16]. The incidence of cardiotoxic effects associated with cancer therapies are 3–26, 7–28, and 2–34 % for anthracycline therapy (Doxorubicin), alkylating agents (Cyclophosphamide), and trastuzumab (Herceptin), respectively [8, 17]. Anthracyclines are associated with cumulative, dose-dependent, irreversible type I cardiac dysfunction [18, 19] as a result of increased reactive oxygen species generation, which causes cardiomyofibril damage, loss, and vacuolization, while trastuzumab-related reduction in left ventricular ejection fraction (LVEF) tends to be reversible (type II dysfunction) [16, 18–20]. The clinical outcomes of anthracycline toxicity can include dilated cardiomyopathy, reduced LVEF (systolic dysfunction), and asymptomatic diastolic dysfunction due to dysregulation in calcium handling, which leads to impaired left ventricular relaxation [21]. Over time, diastolic dysfunction and diminished coronary conductance may be compounded by preexisting comorbidities that reduce coronary flow or increase oxygen demand, such as atherosclerosis or hypertension [21]. Moreover, when these therapeutic agents interact with genetic and health characteristics of the individual, the long-term impact of cancer treatment is difficult to predict.

Radiation for left-sided breast cancer is also associated with an increased risk for CAD, albeit the incidence of cardiac toxicity from radiation therapy used in older trials (before 1980) is higher than that compared with more modern technologies [22]. However, while improvement in radiation techniques have resulted in less radiation delivered to the heart, some individuals with left-sided breast cancer may receive cardiotoxic doses (i.e., that no more than 5 % of the heart’s volume should receive 25 Gy), leaving them vulnerable to an increased risk of CVD development 10 years post treatment [23]. Furthermore, the cardiotoxic effects of anthracycline-based chemotherapy may be exacerbated by radiation therapy as a result of endothelial cell loss and subsequent inflammatory responses [6]. Radiation therapy can also affect the integrity of the vascular system via increased expression of inflammatory gene expression that increases the activity of nuclear factor-kappaB, a transcription factor associated with the development of atherosclerosis [24].

While the toxic effects of cancer therapies on the heart have been well documented [25], evidence of vascular system toxicity is nascent. Animal studies highlight the acute effects on the vasculature of doxorubicin administration, with CVD-related impairments reflected by endothelial cell microthrombus formation and reduced blood flow [26] and increased IMT of the carotid artery [27]. In humans, radiation to the thoracic area and head and neck has been associated with vascular impairments including microthrombi development, vessel occlusion, and increased carotid intima-media thickness (IMT) [28, 29]. Similarly, the deleterious effects of chemotherapy on the vasculature are also coming to light. Increased arterial stiffness, increased carotid artery IMT, and reduced common carotid artery compliance in breast cancer survivors have been observed following 6 months of standard anthracycline therapy [30–33].

Importance of regular cardiovascular health assessments

In cancer survivors, subclinical cardiac dysfunction as a result of cardiotoxic therapies can stabilize and the individual remain asymptomatic [21]. However, the damage persists and it may be some months or years post treatment completion when symptomatic events appear [21]. Recently, recommendations for patient management following radiation therapy include screening echocardiography 5 years after exposure in high-risk individuals and 10 years after exposure in those who remain asymptomatic [6]. Given the high risk of CVD, this screening period is sub-optimal, particularly when considering that the progression of CVD may be asymptomatic. Therefore, it is imperative that clinicians and clinical scientists have at their disposal non-invasive, practical, valid, and reliable techniques for tracking CVD risk.

The following section provides an overview of non-invasive, widely available, and relatively inexpensive techniques for assessing endothelial function, central and regional arterial stiffness, central blood pressures, and carotid IMT. These tools exhibit acceptable reliability and validity, are relatively practical, and can be used for routine serial assessments. Clinical assessment recommendations are also provided.

Assessing cardiovascular health

Endothelial function

Functionally, the endothelium is a large autocrine, paracrine, and endocrine organ that plays a key role in vascular homeostasis [34]. Endothelial dysfunction is a pivotal, yet potentially reversible step that has been shown to precede and predict overt CVD [35]. The established methodologies for evaluating peripheral endothelial function include strain-gauge venous occlusion forearm plethysmography [36], ultrasound measured flow-mediated dilation (FMD) [36, 37], peripheral arterial tonometry (e.g., using the EndoPat device) [36], and laser Doppler flowmetry [36]. These techniques assess the vasodilatory responses to endothelium-dependent stimuli such as acetylcholine and increased shear stress, and to endothelium-independent stimuli, including sodium nitroprusside and glyceryl trinitrate. The FMD test is considered the “gold standard” for assessing endothelial function [36–38]. FMD is a non-invasive, valid [39], reliable [40, 41], and relatively inexpensive technique, but is highly technical and requires a skilled operator [37].

The FMD test is conducted using commercial duplex Doppler ultrasound, typically on the brachial artery [36, 37]. FMD is expressed as the percentage increase in brachial artery diameter above baseline, imaged using B-mode, following 5-min of down-stream ischemia, induced using a tourniquet inflated to supra-systolic pressure. The stimulus for FMD, i.e., blood flow-induced elevation in shear stress following ischemia, should be recorded simultaneously, to ensure that change in FMD over time can be attributed to altered endothelial function, and not a change in the stimulus [42]. Dedicated image analysis software is required to ensure objective and reliable assessments of change in diameter [43]. Commercial software is available (e.g., Vascular Tools 5, Medical Imaging Applications LLC) for semi-automated assessments of change in diameter and shear stress. Actual analysis time is dependent on the quality of recorded images.

The prognostic strength of FMD was summarized in a meta-analysis conducted by Inaba et al. [39], which included 4 population-based cohort studies and 10 convenience-cohort studies (n = 5.547); adjusting for confounding risk factors, a 1 % increase in brachial artery FMD equated to a 13 % (95 % CI 9–17 %) reduction in the risk of future cardiovascular events. Moreover, it should be recognized that the studies included in the meta-analysis [39] dated back to 2000; ultrasound technology and image analysis routines have subsequently advanced, suggesting the estimated sensitivity of FMD to be conservative. A limited number of studies have used FMD on women with breast cancer, reporting positive effects of granulocyte colony-stimulating factor [44] and tamoxifen [45, 46] treatments.

Arterial stiffness

Arterial stiffness is a general term that collectively describes distensiblility, compliance, and elastic modulus of the arterial vascular system. Arterial stiffness can be measured locally, centrally (systemically), or regionally. Local measurements provide important physiological information and are more quantitative and sensitive than systemic indices [47–50]. However, local assessments provide no indication of how the artery of interest interacts with central function (i.e., the heart) as part of an integrative system. Central arterial stiffness affects the global buffering properties of the arterial system; just as arterial blood pressure can be considered a global value of hemodynamic load, central arterial stiffness reflects the overall opposition of large arteries to the pulsatile effects of ventricular ejection [47–50]. Regional arterial stiffness is measured at arterial sites of major physiologic importance such as the aorta where the arterial buffering function is principally expressed, or on a particular limb [47–50].

A number of methodologies have been applied to the in vivo assessment of arterial stiffness. These methodologies fall into three broad groups: (1) relating change in area of an artery to distending pressure, i.e., local arterial stiffness; (2) measuring pulse wave velocity (PWV), i.e., regional arterial stiffness; and (3) pulse wave analysis (PWA), i.e., central arterial stiffness (augmentation index, AIx). Ultrasound and magnetic resonance imaging (MRI) are capable of measuring local arterial stiffness [48, 51] as well as PWV [52, 53], but these methodologies require expensive equipment (especially in the case of MRI), a high level of technical expertise, and are often impractical within the clinical or epidemiological setting. PWV can also be assessed using dedicated equipment, including oscillometric, tonometric, volume plethysmographic, and photoplethysmographic devices, of which applanation tonometry is considered the gold standard [49].

The PWV is the speed at which the forward pressure wave is transmitted through the vascular tree [39, 54]. PWV is calculated by measuring the time taken for the arterial waveform to pass between two points a measured distance apart, and involves taking readings from the two sites simultaneously, or gating separate recordings using ECG [49]. Carotid-femoral PWV represents the aortoilliac pathway and is considered the “gold standard” measurement of arterial stiffness [49]. The absolute references values have been defined for carotid-femoral PWV, and are presented with respect to gender, age, and blood pressure [55]. Commercial devices (e.g., SphygmoCor, AtCor Medical) will automate PWV calculations and compare the outcome to reference values. A meta-analysis [56] of 17 longitudinal studies (n = 15,877, mean follow-up 7.7 years) reported that for an increase in carotid-femoral PWV by 1 m/s, the risk of future cardiovascular events, cardiovascular mortality, and all-cause mortality increased by 14, 15, and 15 %, respectively. Recently, Drafts et al. [31] investigated the effects of 6 months low to moderate dose anthracycline-based chemotherapy in 53 men and women with breast cancer, leukemia, or lymphoma and reported that aortic PWV increased from 6.7 ± 0.5 to 10.1 ± 1 m/s (P = 0.0006), confirming a previous a report by Chaosuwannakit et al. [30].

The technologies applied to PWV, as discussed above, have also been employed for PWA to allow estimation of AIx (central arterial stiffness). While applanation tonometry is considered the gold standard for conducting PWA [49], values obtained from oscillometric devices closely align with those obtained from tonometry [54, 57], and oscillometry presents a simple, observer-independent, and reliable [57] technique that is suitable for use within clinical settings. Typically, the periphery pressure waveform is monitored on the radial artery [54] at the level of the wrist with tonometry, or on the brachial artery using oscillometry [54], and a generalized transfer factor is used to generate the corresponding aortic arterial waveform [54, 57]. AIx is typically calculated by expressing augmentation pressure (AP) as a percentage of central pulse pressure [39, 54], where AP is the additional aortic systolic pressure generated by the return of the reflected waves at the central aorta [39, 54]. Commercial devices (e.g., SphygmoCor, AtCor Medical or Cardioscope, Pulsecor) will automate AIx calculations and compare the outcome to reference values, with respect to age and gender. Increased AIx has been significantly correlated to the degree of CAD [58] and demonstrated to independently predict cardiovascular risk and mortality [56, 59, 60]. A meta-analysis [59] of 11 longitudinal studies (n = 5,648, mean follow-up 45 m) reported that for a 10 % increase in AIx the risk of future cardiovascular events and all-cause mortality increased by 32 and 38 %, respectively. To the best of our knowledge, there is no literature reporting use of AIx in breast cancer, although raised AIx has been reported in men with prostate cancer following induced hypogonadism [61].

Central blood pressures

Central blood pressures can be derived from PWA, as described above. Tonometric [62] and oscillometric [63, 64] derived central blood pressures are highly valid, demonstrating excellent agreement with actual pressures derived from simultaneous aortic catheter assessments. The prognostic value of central blood pressures has been recognized by expert consensus [65, 66] and a meta-analysis published in 2010 [59] concluded that central pulse pressures were a stronger determinant of cardiovascular events than peripheral pulse pressures, although this was only of borderline significance (relative risk 1.32 vs. 1.19, P = 0.057). A more recent study [67] reported that monitoring central blood pressures, as opposed to conventional peripheral blood pressures, aided in the management of hypertension, leading to decreased medication use without adverse effects of left ventricular mass. To the best of our knowledge, use of central blood pressures has not been reported in patients with breast cancers, although induced hypogonadism has been reported to have no effect in men with prostate cancer [61].

Intima-media thickness (IMT)

Over the past two decades B-mode ultrasound has become a popular technique for determining the anatomic location and progression of atherosclerosis in the carotid artery [48, 68, 69]. A range of protocols exist, including imaging the common carotid artery, the carotid bifurcation, the internal carotid, as well as imaging both the left and right carotid artery [69]. The most frequently reported IMT measure is an average of the far wall of the common carotid artery from both the right and left sides [69]. The measurement protocol may considerably influence the carotid IMT value reported; however, when a standardized protocol is coupled with automated image analysis software (e.g., Vascular Tools 5, Medical Imaging Applications LLC), carotid IMT assessments are highly objective and reliable [68]. Carotid IMT assessments are safe, inexpensive, quick, and well tolerated by the patient.

In its early stages, atherosclerosis is restricted to the intimal layer of the arterial vessel wall. Ultrasound imaging cannot discriminate between the intimal layer and the medial layer of the vessel wall. Thus, an increased carotid IMT may reflect either increased intimal thickening, increased thickening of the medial layer, or a combination of both. Nonetheless, IMT assessments, which have been endorsed by the American Heart Association [70], are related to traditional CVD risk factors and atherosclerosis elsewhere in the arterial system [71], and demonstrate a consistent and gradual relation to risk of vascular events [72]. However, despite this continuous relationship between CIMT and CVD risk, an absolute definition of an abnormally high CIMT is problematic due to the strong influence of age on arterial wall thickness [73]; therefore, age-specific reference values are recommended [73, 74].

The Kuopio Ischaemic Heart Disease Risk Factor Study found that myocardial infarction risk increased by 11 % for each 0.1 mm increase in carotid IMT [75]. This has been verified by a more recent meta-analysis [72], which included 8 general population studies (n = 37,197, mean follow-up 5.5 years), and reported that for an absolute carotid IMT increase of 0.1 mm, the future risk of a myocardial infarction increased by 10–15 %. A limited number of studies have reported the use of carotid IMT on women with breast cancer, including positive effects of tamoxifen treatment [45, 46] and negative effects of chemotherapy [32, 33, 76]. Recently, Mizia-Stec et al. [33] monitored 36 patients (35–68 years) pre- and post-anthracycline treatment, and reported an increase in carotid IMT from 0.59 ± 0.1 to 0.74 ± 0.11 mm (P = 0.001), supporting previous findings by Kalabova et al. [32].

Clinical assessment recommendations

A summary of recommended vascular health screening tools are listed in Table 1. The tools have been listed in order of preference, based on a trade-off between suitability for use in clinical practice and validity. Carotid-femoral PWV, the gold standard assessment of arterial stiffness [49], is not quite as simple to conduct as PWA (AIx and central blood pressures), but is a simpler process than carotid IMT or FMD, and commercial devices (e.g., SphygmoCor) will automate the calculations and provide comparisons to reference values.

Table 1 Recommended non-invasive vascular health screening tools
Full size table

The optimal vascular health screen exam will include more than one outcome. Each of the highlighted vascular health screening tools may reflect distinct aspects of the atherosclerotic process, and may confer additive prognostic information. In support, a number of studies have reported poor to modest correlations among FMD, IMT, and PWV [77–79]. Holewijn et al. [80] assessed the additive prognostic capacity of FMD, IMT, PWV, AIx, AP, central systolic pressure, ankle-brachial index, and carotid plaque in 1,367 middle-aged, CVD-free patients. Cardiovascular events were validated after a mean follow-up of 3.8 years. In women, cardiovascular risk stratification improved by adding combinations of non-invasive outcomes, but in men risk stratification only improved for men at intermediate risk [80]. The authors concluded that, at least in the middle-aged, risk stratification using traditional cardiovascular risk factors works reasonably well in men, but fails in women [80].

A number of devices are available for recording more than one outcome. High-resolution portable duplex Doppler ultrasound devices (e.g., T3200, Terason) are capable of assessing carotid IMT as well as FMD. Oscillometric PWA devices are available (e.g., Cardioscope, Pulsecor) that provide automated assessments of AIx and central blood pressures, taking approximately 1 min to complete a recording (excluding preparation time). The recently released SphygmoCor XCEL (AtCor Medical) records AIx and central blood pressures using oscillometry on the upper arm, and carotid-femoral PWV by combing oscillometry on the upper leg with tonometry on the common carotid artery. An example of vascular health screening routine is depicted in Fig. 1. The suggested routine will allow non-invasive estimates of different morphological and functional abnormalities of the cardiovascular system, which when coupled with published reference values will allow the clinician to make better informed treatment decisions. The optimal routine will follow standard testing protocols, as well as standard patient preparation protocols, including: reporting to the clinic following an overnight fast, resting quietly in the supine position for 20 min prior to testing, testing in the supine position, taking serial measurements at the same time of day, and if pre-menopause, testing between days 1 and 7 of the menstrual cycle [37, 81].

Fig. 1
figure 1

Example vascular health screening routine. Measurements include: augmentation index (AIx) and central blood pressures (CBP) using pulse wave analysis (PWA) and pulse wave velocity (PWV)

Full size image

Recently, Carver et al. [82] presented a paradigm for screening cancer survivors exposed to cardiotoxic therapies, with post treatment screening beginning at the 5 year mark. Recommended assessments include clinical history, physical examination, blood biomarkers (B-type natriuretic peptide, lipids), echocardiogram (or radionuclide angiogram), and an electrocardiogram [82]. Future screening intervals are then based on the results of the findings. We would recommend assessment of vascular function be included in this screening protocol.

Conclusion

Given the long latency periods between cancer treatment completion and potential symptomatic CVD, there is a need to exploit assessment protocols that detect vascular changes before symptoms appear. Flow-mediated dilation, central and regional arterial stiffness, central blood pressures, and IMT are non-invasive, widely available, and relatively inexpensive techniques that have acceptable reliability and validity. Long-term, further research is warranted to establish independent and additive reference values for these methods in patients with breast cancer. Short-term, there is sufficient evidence to encourage the use of these techniques as part of routine serial assessments, and to help guide treatment strategies.