Vasomotor Symptoms and Cardiovascular Health: Findings from SWAN and the MsHeart/MsBrain Studies

Abstract

Vasomotor symptoms (VMS) are often considered the classic menopausal symptom and are experienced by most women at during the menopause transition. VMS are well-established to be associated with decrements in quality of life during the menopause. More recent research also links VMS to poorer cardiovascular health. This review summarizes key insights about links between VMS and cardiovascular disease (CVD) risk that come from the Study of Women’s Health Across the Nation (SWAN), a longitudinal epidemiologic cohort study of the menopause transition, as well as from the MsHeart/MsBrain studies, clinical studies that leverage vascular imaging and brain imaging as well as wearable technologies that provide objective indicators of VMS. Using a range of methodologies and extensive consideration of confounders, these studies have shown that frequent and/or persistent VMS are associated with adverse CVD risk factor profiles, poorer underlying peripheral vascular and cerebrovascular health, and elevated risk for clinical CVD events. Collectively, SWAN and the MsHeart/MsBrain studies form complementary epidemiologic and clinical studies that point to the importance of VMS to women’s cardiovascular health during the menopause transition and beyond.

VMS, also known as hot flushes/flashes and night sweats, are the classic symptom of the menopause transition. VMS are characterized by acute feelings of heat accompanied by sweating and flushing that are typically experienced around the face, neck, and chest [1]. Most women will experience VMS during the menopause transition [2], and for a quarter of women, VMS occur daily [3]. Key insights about VMS come from the Study of Women’s Health Across the Nation (SWAN), a longitudinal epidemiologic cohort study that followed 3302 women living in the United States as they transitioned through the menopause, and the MsHeart/MsBrain studies, clinical studies that have leveraged objective assessments of VMS in conjunction with vascular and neuroimaging to understand VMS and their implications for women’s health. This review will highlight key findings from SWAN and the MsHeart/MsBrain studies about VMS and their links to women’s cardiovascular and brain health.

Natural History of VMS

SWAN has provided critical information about the natural history of VMS. SWAN enrolled 3302 pre- or early perimenopausal women across five racial/ethnic groups between the ages of 42 and 52 and followed participants for approximately 25 years [4]. SWAN findings have revealed pronounced racial/ethnic differences in VMS among women living in the United States: Findings range from approximately 50% of Japanese women living in the United States to 80% of African American women reporting VMS, respectively [2]. Further, frequent VMS persist for approximately seven years, and women who started their VMS earlier in the transition had the most persistent VMS [5]. Moreover, SWAN results indicated that women experienced their VMS at different times, showing four trajectories of VMS across the transition: 1) VMS early in the transition, when women are still menstruating; 2) VMS later in the transition, primarily after women’s final menstrual period; 3) VMS in the few years around the final menstrual period; and 4) VMS experienced persistently across the entirety of the transition, from the late reproductive years to well into the postmenopause [6]. These findings parallel those of other studies, including the Australian Longitudinal Study on Women’s Health which found broadly similar trajectories [7].

VMS have implications for quality of life, mental health, and sleep during the menopause transition. For example, in SWAN, VMS were associated with poorer social, emotional, and physical quality of life [8]. VMS were a risk factor for depressed mood [9] as well as reported sleep problems [10] during the menopause transition. However, despite the profound effect VMS can have on women’s quality of life, VMS have generally been regarded as not having implications for physical health. Increasingly, findings have called this notion into question.

Cardiovascular Disease (CVD) Risk Factors

SWAN and other studies have shown VMS associated with adverse CVD risk factor profiles [11]. SWAN findings have indicated that VMS, particularly when experienced frequently, were associated with elevated HOMA scores [12], an indicator of insulin resistance. Further, in SWAN, VMS were associated with adverse lipid profiles, particularly higher LDL cholesterol and triglycerides [13]. VMS in SWAN were also associated with a 40% increased risk of incident pre-hypertension or hypertension over the subsequent decade [14]. In contrast to many other studies that utilized cross sectional designs [11], SWAN is notable in its investigation of these associations in a longitudinal manner, using regular, prospective measurements of VMS and CVD risk factors over time.

VMS and Subclinical CVD

SWAN investigated associations between VMS and subclinical CVD. Subclinical CVD indices use ultrasound or computed tomography imaging of the vasculature to index the health of the vasculature before clinical disease is present and help identify those at risk for clinical CVD. These measures include ultrasound-assessed flow mediated dilation (FMD), an index of endothelial function, and carotid intima media thickness (IMT) and plaque, indicators of carotid atherosclerosis. Further, computed tomography is used to assess calcified plaques in the aorta and the coronary arteries. Among participants in the SWAN Heart ancillary study, women reporting hot flashes in the past two weeks had poorer endothelial function as indexed by brachial artery FMD, as well as greater aortic calcification and coronary artery calcification than women without hot flashes [15]. Associations between hot flashes and FMD or aortic calcification withstood adjustment for demographic factors, CVD risk factors, and endogenous sex hormone concentrations. In addition, in the full SWAN cohort, frequent VMS (experienced six or more days in the prior two weeks) were associated with higher carotid IMT, controlling for demographic factors, CVD risk factors, and endogenous estradiol concentrations [16]. Several other studies did not find associations between reported VMS and subclinical CVD[17–19] notably, most of these studies were conducted among relatively young, healthy women relative to those in SWAN, suggesting that these associations may most apply to women with some baseline CVD risk.

These findings raise the intriguing question about whether VMS are in fact indicating poorer underlying vascular health among midlife women. However, like most epidemiologic cohort studies, SWAN findings are limited by use of brief, questionnaire assessments of VMS. These questionnaires have known memory and reporting biases and provide only broad information about VMS [20]. Questionnaire-assessed sleep VMS are also especially prone to error and are influenced by sleep itself [21]. Thus, to fully understand the nature of the relationships between VMS and CVD risk, use of objective assessments of VMS is critical. Objective assessments of VMS provide validated assessments about the occurrence and timing of VMS during both wake and sleep and avoid the reporting, affective, and memory influences on recalled VMS [21].

The MsHeart study was designed to test whether and by what mechanisms objectively-assessed VMS were associated with indicators of CVD in women [22]. MsHeart enrolled a carefully-screened sample of 304 nonsmoking late perimenopausal and postmenopausal women between ages 40-60 who were free of clinical CVD and not taking known medications (e.g., hormone therapy, SSRI/SNRI antidepressants) that influence VMS. MsHeart incorporated wearable technologies to provide both objective and subjective assessments of both VMS and sleep as well as prospective ecological momentary self-report assessments of VMS; vascular imaging; and sex steroid hormones assayed via liquid chromatography-tandem mass spectrometry (LC-MS/MS). The primary vascular outcome in MsHeart was carotid IMT, as it is associated with clinical CVD beyond CVD risk factors [23] and has predictive utility among midlife women among whom other indicators (e.g., coronary artery calcification [24]) are typically too low to assist in risk stratification. Thus, complementing the broad, epidemiologic perspective provided by SWAN, MsHeart is a clinical study that provides a greater degree of rigor, control, and precision to rigorously test relationships between VMS and CVD risk.

MsHeart found that among women reporting VMS, more frequent objectively-assessed VMS were associated with higher carotid IMT in a dose-response fashion [22]. These associations persisted controlling for demographic factors and standard CVD risk factors such as blood pressure, lipids, or insulin resistance. In fact, among women reporting VMS, the variance in IMT accounted for by the frequency of objective VMS was larger than that of any of the standard CVD risk factors. Further, approximately half of the MsHeart participants showed evidence of carotid plaque, and among women reporting VMS, more frequent VMS were associated with greater carotid plaque. Sex steroid hormones, which were only weakly related to subclinical CVD [25], did not explain relationships between VMS and carotid atherosclerosis. Thus, in MsHeart, more frequent physiologically-assessed VMS were associated with greater subclinical atherosclerosis and these associations were not accounted for by CVD risk factors or by sex hormones.

VMS and Clinical CVD Events

Studies have also examined associations of VMS to clinical CVD events such as heart attacks, strokes, heart failure, and CVD mortality. An early post hoc analysis of over ten thousand women found that night sweats were associated with increased risk of coronary heart disease over the subsequent decade [26]. This study provided important initial evidence, yet was not designed for this purpose and had some key methodologic limitations. Rigorously testing the association between VMS and clinical CVD presents multiple methodologic challenges including the need for repeated prospective assessments of VMS over long periods of time starting at midlife in combination with extended follow up times to characterize CVD events, which occur typically in women’s 60s and 70s. SWAN met these criteria. In SWAN, we examined the association between reported VMS assessed over up to 16 visits in relation to CVD events occurring over the 25-year follow up period. Findings indicated that more frequent VMS reported at baseline were associated with a 50% increased risk of CVD events over the following 25 years [27]. Further, frequent VMS experienced persistently over the menopause transition – for at least of third of attended visits with frequent VMS – was associated with a 77% increased risk of subsequent CVD events. These models adjusted for a range of covariates including CVD risk factors, and associations persisted when additionally adjusting for endogenous estradiol concentrations. Conversely, the Australian Longitudinal Study on Women’s Health provided mixed findings, with one report showing associations between VMS and later coronary heart disease events and another subsequent report showing no significant association between VMS and incident CVD [28,29]. Another notable study from the INTERLACE consortium found that severe VMS were associated with a doubling of risk of CVD [30]. Finally, a recent meta-analysis concluded that associations between VMS and subsequent CVD events are primarily observed among women who are aged under 60 years at the baseline examination [31], understandable given the ages at which VMS are most prevalent. Collectively, although findings are not universal, they generally find frequent, severe, or persistent VMS associated with increased risk for later CVD events.

Mechanisms Linking VMS to CVD Risk

A range of potential mechanisms have been considered in linking VMS to CVD risk. As women with more VMS also have more adverse CVD risk factor profiles, CVD risk factors are important to consider. However, associations between VMS and subclinical or clinical CVD persist adjusting for CVD risk factors. Another key potential mechanism is changes in endogenous estrogens, particularly estradiol, which is often implicated in menopause-related increases in CVD risk. However, associations between VMS and CVD risk typically persist adjusting for endogenous steroid hormone concentrations, including in MsHeart, which measured estrone and estradiol via LC-MS/MS, a method able to characterize the very low levels of estradiol that characterize the postmenopause. Sleep is also a critical pathway to consider in links between VMS and CVD risk, as poor sleep is an increasingly well-established risk factor for CVD [32,33]. Whereas some research has demonstrated the links between VMS and subclinical CVD or cerebrovascular risk to be independent of and potentially synergistic with VMS [34,35], ongoing research is warranted to understand the complex inter-relationships between VMS, sleep, and CVD risk. Other mechanisms that have been considered in relationships between VMS and CVD risk include inflammatory factors, adipokines, or clotting factors, which we and others have found altered among women with VMS [36–38]. However, these factors were considered carefully in MsHeart and were not explanatory. Finally, MsHeart implemented ambulatory 24-hour electrocardiogram to characterize autonomic nervous system control over heart rate (heart rate variability) in conjunction with objective VMS monitoring. Findings indicated that VMS were acutely associated with reduced respiratory sinus arrhythmia (RSA), indicating acute decreases in parasympathetic nervous system influence over heart rate during objectively-documented VMS [39], a profile linked to CVD risk. These decreases were observed irrespective of whether the VMS were reported or perceived. However, in MsHeart, these autonomic nervous system alterations did not explain associations between VMS and carotid atherosclerosis. Other pathways such as alterations in the hypothalamic pituitary adrenal axis [40], epigenetic changes [41], or mechanisms informed by recent advances in the understanding of neurobiology of VMS [42] warrant consideration in future work.

Timing of VMS in Relation to CVD risk

Several studies have considered whether the timing of VMS is germane to their relationship to CVD. Some findings from the Women’s Health Initiative Observational Study suggested it was later occurring VMS (at the time of enrollment) as opposed to earlier VMS (recalled as occurring around the time of menopause onset) that were most important for CVD risk [43]. Our findings from the Women’s Ischemia Syndrome Evaluation (WISE) study cohort indicated that it was women who had VMS that began yearly (before age 42) that had poorer endothelial function and higher CVD mortality relative to women who had VMS that began later in life [44]. Finally, the INTERLACE collaborative found both early and late VMS associated with CVD events [30]. However, many of these studies were limited by a range of factors including their brief measures of VMS recalled over long periods of time subject to a high degree of error.

The SWAN and MsHeart studies considered whether the timing of VMS was related to subclinical CVD. First, in SWAN we considered the four trajectories of VMS: few or no VMS over the transition, VMS occurring early in the transition, VMS occurring later in the transition, and VMS persisting throughout the entire transition. In minimally adjusted models, two trajectories emerged as related to higher IMT: VMS occurring early in the transition and VMS persisting over the entire transition. However, after adjustment for race/ethnicity and CVD risk factors, only VMS occurring early in the transition were related to higher IMT [45]. In MsHeart we also considered the timing of VMS in relation to subclinical CVD. Whereas age did not modify the relationship between VMS and IMT, age did modify the relationship between VMS and FMD, such that objectively-assessed VMS were associated with poorer endothelial function only among the younger women in the cohort [46]. Further, we considered CVD events in SWAN, finding that it was VMS reported at baseline (early in the transition) or persistently throughout the transition that were linked to elevated CVD risk after adjustment for CVD risk factors [27]. Thus, findings from SWAN and MsHeart suggest that it is the women experiencing VMS early in the transition or persistently over the transition that may be at greatest cardiovascular disease risk. Moreover, findings from a recent meta-analysis on VMS and clinical CVD suggest that the that associations between VMS and subsequent CVD events are primarily observed among women who are aged under 60 years at the baseline examination [31], However, findings of age or timing effects in relations between VMS and CVD risk require ongoing investigation and replication.

VMS and the Cerebrovasculature

The close inter-connections between the cardiovascular system and the brain are increasingly appreciated, with CVD and its risk factors well established risk factors for cognitive decline and dementia [47]. Moreover, the menopause transition is not only a time of accelerated vascular risk [48], but also a time of increased cognitive complaints (e.g., brain fog) and decrements in memory performance [49]. Work conducted by Maki and colleagues found that more frequent VMS, particularly objectively-assessed VMS occurring during sleep, have been associated with poorer memory performance in women [50]. Integrating these lines of research, we together conducted pilot research, asking 20 of the MsHeart participants to undergo neuroimaging to quantify brain white matter hyperintensities (WMHs), or indicators of small vessel disease in the brain linked to future stroke, dementia, and mortality [51,52]. Among these women, we found that VMS, particularly when occurring during sleep, were associated with more WMHs, and that these associations were not explained by CVD risk factors, sleep, or estradiol [53]. These findings raised the possibility that the links between VMS and the peripheral vasculature extended to the cerebrovasculature. However, as this was a small pilot study, these findings could only be regarded as preliminary.

MsBrain Study was designed to investigate associations between VMS and neurocognitive health in women, while also considering the role of cardiovascular health, sleep, and sex hormones in these associations. Conducted five years after MsHeart, MsBrain enrolled 230 MsHeart participants and community participants who underwent both objective VMS and sleep monitoring, neuropsychological testing, vascular imaging, and neuroimaging. Among 226 MsBrain participants, we found that objectively-assessed VMS during sleep were associated with greater WMHs in the brain [34]. These associations were not explained by sleep, by CVD risk factors, nor by endogenous estrogens. Demonstrating the connections between the peripheral and cerebrovasculature, women with higher IMT did have greater WMHs in MsBrain [54]; however, IMT did not fully explain relationships between VMS and WMHs [34]. Thus, bridging the heart and brain, MsHeart and MsBrain together indicate that VMS are not only associated with poorer vascular health in the periphery, but also in the brain.

Nature of Relationships Between VMS and CVD risk

Accumulating data reveal consistent links between VMS and cardiovascular health. However, the precise nature of these associations, and particularly any causal associations between VMS and cardiovascular health, is unknown. It is possible that adverse underlying vascular health gives rise to both VMS and CVD or that a third yet unidentified factor gives rise to both VMS and CVD. In the case of cognition, there is indication that the degree of VMS improvement in response to a novel treatment (stellate ganglion block) correlated systematically with improvements in verbal learning [55], suggesting a potential causal set of relations. However, more research, particularly experimental research, is required to understand the precise nature of relationships between VMS, cardiovascular health, and neurocognitive health. Collectively, the available data show that women with frequent VMS are at risk for poor cardiovascular outcomes, yet any causal relationship between VMS and CVD is as yet unclear.

Summary

Much has been learned about VMS and their importance to women’s health. Two studies that have shed important light on VMS and their implications for health among women in the United States include SWAN, a 25-year epidemiologic cohort study of the menopause transition, and the MsHeart/MsBrain studies, clinical studies leveraging physiologic VMS monitoring and vascular and neuroimaging to address questions about VMS and their implications for health. Together, they form complementary epidemiologic and clinical studies that have yielded key findings about VMS and their links to cardiovascular and brain health. These studies have shown VMS linked to adverse CVD risk factors, including hypertension, insulin resistance, and poorer lipid profiles, greater subclinical CVD, including poorer endothelial function and greater carotid atherosclerosis, as well as greater risk for clinical CVD, such as myocardial infarction, stroke, and CVD mortality later in life. VMS, particularly when experienced during sleep, are linked to poorer cerebrovascular health. Future research is required to elucidate the underlying mechanisms as well as any causal nature of associations between VMS and health. However, VMS, formerly regarded as short, time-delimited incidental symptoms of menopause, are increasingly understood to be of potential importance to women’s cardiovascular health.