Menopausal Hot Flashes and White Matter Hyperintensities

Introduction

Hot flashes are the cardinal symptom of the menopausal transition, experienced by over 70% of women during the menopausal transition.¹ They have long been understood to be associated with reductions across all domains of quality of life.² However, hot flashes have recently been shown to be associated with decreased physical health.

Several studies have shown hot flashes to be associated with elevated subclinical cardiovascular disease (CVD), including poorer endothelial function, higher intima media thickness, and aortic calcification.^{3, 4} Hot flashes have also been linked to an adverse CVD risk factor profile, including elevated blood pressure,⁵ higher lipids,⁶ a more insulin resistant profile,⁷ and a more pro-coagulant hemostatic profile.⁸ Further, hot flashes, particularly those measured physiologically, have been associated with decreased cardiac vagal control.⁹ Notably, associations between hot flashes and CVD risk indices persist after adjusting for other standard CVD risk factors.

The existing literature on hot flashes has largely focused on changes in peripheral physiology. Although hot flashes are conceptualized as largely a central nervous system event,¹⁰ very few studies have examined hot flashes in relation to brain health. Elevated traditional and novel CVD risk factors, such as hypertension, insulin, and autonomic nervous system dysfunction have been associated with adverse changes in the brain, notably lesions in the white matter.^11-13 These lesions appear as white matter hyperintensities (WMH) on T2-weighted magnetic resonance imaging (MRI) scans and are thought to develop due to small vessel disease in the brain.¹⁴ WMH are linked to later stroke, cognitive decline and dementia, and mortality.¹⁵ No studies have examined whether hot flashes are linked to WMH in the brain.

We tested whether wake and sleep hot flashes were associated with WMH. We tested these hypotheses in a sample of midlife women who underwent both brain imaging and detailed ambulatory diary and physiologic hot flash monitoring. We considered the role of a range of potential explanatory factors and confounders in observed associations.

Materials and Methods

Results

Participants were on average 55 years old and approximately a quarter of the women were African American, with the remainder non-Hispanic White. The participants were on average normotensive, overweight and postmenopausal. Nine women reported having hot flashes, yet on physiologic monitoring 14 women showed physiologic hot flashes, a finding that is consistent with prior findings that women tend to under-report hot flashes relative to physiologic monitoring.²⁹ For the sample as a whole, women reported an average of 3 hot flashes/24 hours, and showed 8 hot flashes/24 hours on physiologic monitoring.

When considered categorically, women with hot flashes (physiologically-verified) had marginally higher WMH [relative to women without hot flashes: beta, b (standard error, SE)=0.0009 (0.0004), p=0.07], controlling for age, race, and BMI. Considered continuously, more physiologic hot flashes, particularly during sleep, were associated with significantly greater WMH. (Table 2; Figure 1). Notably, in these models the magnitude of the effect size for sleep hot flashes [for each additional hot flash: b(SE)=0.0002 (0.0001), p=0.03] was larger than that for age [per year: b(SE)=0.000007 (0.00005), p=0.90] or race [White vs African American: b(SE)=0. 0005 (0.0005), p=0.31], and comparable to that of BMI [per unit increase in BMI: b(SE)=0.0002 (0.00004), p=0.001], the only other significant variable in the model. No significant associations were observed between self-reported hot flashes and WMH [24-hour self-reported hot flashes, per additional hot flash: b(SE)=-0.000013 (0.00006), p=0.83; self-reported hot flashes during waking hours, per additional hot flash: b(SE)=-0.000017 (0.0001), p=0.86; self-reported hot flashes during sleep, per additional hot flash: b(SE)=-0.00001 (0.0001), p=0.91, controlling for age, race, BMI].

Because findings were observed for hot flashes occurring during sleep, we also adjusted for sleep characteristics, including actigraphy-assessed sleep time, waking after sleep onset, sleep efficiency, and fragmentation; and questionnaire-assessed sleep apnea, sleep quality, or daytime sleepiness. Only higher daytime sleepiness was related to more WMH [per increase in Epworth scale score: b(SE)=0.0001 (0.00004), p=0.02, adjusted for age race, and BMI]. However, controlling for sleepiness in relations between overnight physiologic hot flashes and WMH had little impact on associations [per additional hot flash: b(SE)=0.0003(.00008) p=0.008, adjusted for age, race, BMI, sleepiness].

We performed several additional analyses to determine whether other potential covariates that might relate to hot flashes and might modify the relationship between hot flashes and WMH. These covariates included education, SBP, DBP, menopause stage, lipids, HOMA, blood pressure-lowering medication use, prior hormone therapy use, self-rated health, depressive symptoms, IL6, and hsCRP. Higher LDL cholesterol [per unit increase in LDL: b(SE)=0.00003 (0.000008), p=0.01] and higher hsCRP_log [per unit increase in hsCRP_log: b(SE)=0.0005 (0.0002), p=0.01] was associated with more WMH (both models adjusted for age, race, BMI, sleep physiologic hot flashes). None of the other factors (education, SBP, DBP, menopause stage, HDL, triglycerides, HOMA, blood pressure-lowering medication use, prior hormone therapy use, self-rated health, depressive symptoms, IL6) were significantly related to WMH. Further, inclusion of any these factors (including LDL and hsCRP) in models between sleep physiologic hot flashes and WMH did not substantially alter findings (data not shown). Further, we considered E2_log concentrations in relations between sleep physiologic hot flashes and WMH in models adjusted for age, race, and BMI; relations between sleep physiologic hot flashes and WMH persisted [per additional hot flash: b(SE)=0.0002 (0.0001), p<0.05].

Discussion

This study was the first to investigate the relation between menopausal hot flashes and WMH. We showed that more physiologically-monitored hot flashes during sleep were associated with greater WMH burden among midlife women free of clinical cardiovascular or cerebrovascular disease. These associations were not accounted for demographic factors, CVD risk factors, sleep, E2, or other potential confounding factors. These findings demonstrate a link between menopausal hot flashes and brain health, specifically structural abnormalities associated with ischemic disease.

Some studies have linked hot flashes to elevated subclinical CVD as well as a more adverse CVD risk factor profile.^3-8 Notably, a more adverse CVD risk profile, particularly hypertension and diabetes, has been associated with more WMH which may reflect small vessel disease in the brain.^{11, 12} Greater WMH burden in turn has been linked to elevated risk of stroke, dementia, and mortality.¹⁵ Thus, these data suggest that the adverse CVD profile observed in the periphery in some work may extend to the brain. Importantly, the direction of relations cannot be assumed here - hot flashes may serve as a marker for underlying white matter changes, they may be etiologically involved in these changes in women, or the relations between WMH and hot flashes may be explained by a third factor. The precise nature of relations between hot flashes and WMH should be investigated in future work, including in longitudinal studies.

Prior work has considered other reproductive and hormonal factors, such as hormone therapy use, pregnancy, and preeclampsia in relation to WMH.^30-32 However, no prior work has considered menopausal hot flashes in relation to WMH. WMH can be observed among relatively healthy midlife women, and some evidence indicates that WMH may be particularly related to novel CVD risk factors as compared to traditional CVD risk factors in this group.³³ Further, prior work has demonstrated that reproductive factors, such as bilateral oophorectomy during the pre or perimenopause (which typically induces hot flashes), has been linked elevated risk for cognitive impairment, dementia, and parkinsonism.^{34, 35} Midlife reproductive phenomena such as hot flashes may be particularly relevant to consider in relation to later brain health. Future work should extend the present findings to a wider range of clinical and functional (i.e. cognitive) outcomes as well as to other populations, such as women undergoing oophorectomy.

Findings were observed primarily for physiologically-assessed hot flashes. Physiologically-monitored hot flashes differ from self-reported hot flashes in several important ways. Physiologic hot flash measures have the advantage of not relying upon attention, perception, emotional influences, or adherence to hot flash reporting protocols³⁶. These issues are particular relevant in the ambulatory setting which presents many distractions and competing activities. Thus, typically more hot flashes are detected than are reported in the ambulatory setting.^{9, 29, 37} Physiologic hot flash measures may be particularly useful for measuring sleep hot flashes, when reporting may be particularly difficult and impacted by the quality of sleep itself.³⁸ Finally, as compared to self-reported hot flashes, physiologically-monitored hot flashes have been shown to be more strongly related to other indices such as cardiac vagal control⁹ and cognitive indices such as verbal memory.³⁷

Findings were most pronounced for physiologically-monitored hot flashes detected during sleep. These relations did not appear to be accounted for by sleep characteristics assessed both by actigraphy and by questionnaire. In fact, only daytime sleepiness was related to greater WMH burden, and sleepiness did not account for the observed associations. Notably, sleep apnea in other work has been related to WMH,³⁹ but was not related to WMH here, and sleep apnea is not typically related to hot flashes.⁴⁰ Although sleep was rigorously assessed here, without polysomnography measures of sleep, a role for sleep cannot be definitively ruled out. However, the evidence here does not support a major role for sleep in these associations. It is also notable that relations between physiologic hot flashes and other relevant outcomes, such as cardiac vagal control ⁹ and cognition,³⁷ appear most pronounced for sleep hot flashes. Whether sleep hot flashes and waking hot flashes have different underlying physiologies is not known, and further investigation of the unique relation of sleep hot flashes to WMH is warranted.

The mechanisms that may link hot flashes to WMH are not entirely clear, yet the findings here were not altered when controlling for CVD risk factors, such as blood pressure, BMI, or lipids. Inflammation has been tied to WMH⁴¹ and possibly hot flashes.⁴² We considered IL6 and CRP here, and although CRP was related to WMH, neither marker explained the association between hot flashes and WMH. Further, hot flashes typically occur in the context of E2 withdrawal,¹⁰ which may have implications for brain health.^{32, 43} However, findings persisted controlling for E2, which was measured via gold standard methods that are sensitive to the low levels of E2 observed in postmenopausal women. Notably, as the timing of E2 withdrawal may be relevant to brain health,⁴³ future work with larger sample sizes should consider variations in the timing of hot flashes in relation to WMH. Other work has shown autonomic nervous system profile characterized by increased sympathetic and/or decreased vagal control of the heart associated with hot flashes⁹ and WMH,¹³ which may be a candidate pathway linking these two phenomena. Future work should further consider the mechanisms by which hot flashes may be associated with WMH.

This study has several limitations. The main limitation of the study was its small size, which may have limited power to detect associations. It is notable that such consistent findings were noted between WMH and hot flashes despite this small sample, and findings should be replicated in future work. The majority of women included here were postmenopausal and thereby any differences by menopausal stage could not be investigated. Although extensive sleep indices were included, future work should examine the role of sleep in these relations using additional measures such as polysomnography.

This study had several strengths. This is the first study to investigate menopausal hot flashes in relation to WMH. Hot flashes were investigated via both physiologic monitoring and prospective diary report and during wake and sleep as a woman went about her daily activities. Women completed a detailed brain imaging procedure. Moreover, other related important indices, including sleep, depression, CVD risk factors, and rigorously-measured endogenous E2 concentrations were assessed and their roles considered.

Conclusions

More menopausal hot flashes, particularly hot flashes occurring during sleep hours, were associated with greater WMH burden among midlife women free of clinical cardiovascular or cerebrovascular disease. This study is the first to investigate hot flashes in relation to WMH, showing that changes in white matter can occur early in life among women with hot flashes during sleep. These findings underscore the importance of continued investigation of the brain in investigating the propensity towards, underlying physiology of, and sequelae of hot flashes.