Authors: Catherine Kim (1Departments of Medicine, Obstetrics & Gynecology, and Epidemiology, University of Michigan, Ann Arbor, Michigan;), Pamela J. Schreiner (2Division of Epidemiology & Community Health, University of Minnesota, Minneapolis, Minnesota;), Zhe Yin (3Department of Biostatistics, MD Anderson Cancer Center, Houston, TX;), Rachael Whitney (4Department of Medicine, University of Michigan, Ann Arbor, Michigan;), Stephen Sidney (5Division of Research, Kaiser Permanente Northern California, Oakland, CA;), Imo Ebong (6Department of Medicine, University of California, Davis, CA), Deborah A. Levine (4Department of Medicine, University of Michigan, Ann Arbor, Michigan;)
Categories: Article, vasomotor symptoms, hot flashes, night sweats, menopause, stroke, cardiovascular disease
Source: Menopause (New York, N.Y.)
Authors: Catherine Kim, Pamela J. Schreiner, Zhe Yin, Rachael Whitney, Stephen Sidney, Imo Ebong, Deborah A. Levine
To examine whether vasomotor symptoms (VMS) and migraine headaches, hypothesized to be vasoactive conditions, are associated with greater risk for cardiovascular disease (CVD) events including strokes.
We performed a secondary data analysis of a subset of women (n=1954) in the Coronary Artery Risk Development in Young Adults (CARDIA) study, a population-based cohort, that began data collection at 18–30 years of age. We examined whether migraine headaches and VMS trajectories (characterized as minimal, increasing, and persistent) at CARDIA Exam Year 15 were associated with higher risk of CVD events and stroke (both ischemic and hemorrhagic) using Cox proportional hazards regression models and adjustment for traditional CVD risk factors (age, cigarette use and levels of systolic and diastolic blood pressure, fasting glucose, high- and low-density cholesterol and triglycerides) and reproductive factors.
Among women with minimal VMS (n=835), increasing VMS (n=521), and persistent VMS (n=598), there were 81 incident CVD events including 42 strokes. Women with histories of migraine and persistent VMS had greater risk of CVD (HR 2.25 (95% CI 1.15 – 4.38) after adjustment for age, race, estrogen use, oophorectomy, and hysterectomy compared to women without migraine histories and with minimal/increasing VMS. After adjustment for CVD risk factors, these associations were attenuated (HR 1.51, 95% CI 0.73 – 3.10). Similarly, women with histories of migraine and persistent VMS had greater risk of stroke (HR 3.15, 95% CI 1.35 – 7.34) but these associations were attenuated after adjustment for CVD risk factors (HR 1.70, 95% CI 0.66 – 4.38).
Migraines and persistent VMS jointly associate with greater risk for CVD and stroke although risk is attenuated with adjustment for traditional CVD risk factors.
Vasomotor symptoms (VMS) are a risk factor for cardiovascular disease (CVD) in women, in part because women with more severe and frequent VMS often have poorer CVD risk factor profiles.^1–4^ Migraines, particularly with aura, are also associated with increased risk for CVD,^5–9^ including hemorrhagic^8^ and ischemic stroke,^9^ as well as for VMS.^10^ Both VMS and migraines have each been hypothesized to represent increased vasoreactivity that could potentially predispose women to CVD events. However, it is unclear whether VMS and migraines are risk factors for CVD and stroke independent of traditional CVD risk factors (cigarette use, levels of lipids, blood pressure, and glucose). It is also unclear whether VMS and migraines are associated with CVD and stroke independent of exogenous estrogen use, since exogenous estrogen use in women with focal migraines may increase stroke risk,^11^ and exogenous estrogen use may be prescribed for treatment of VMS^12^ and in women with histories of gynecologic surgery, particularly oophorectomy.^13^
The Coronary Artery Risk Development in Young Adults (CARDIA) study is a population-based study that collected CVD risk factors every 5 years beginning when participants were aged 18–30 years of age until the most recent exam when women were aged approximately 61 years. CARDIA also has collected migraine history before and at the Year 15 Exam (Y15) exam when women were approximately 40 years of age and every 5 years thereafter. Using these data, we examined the association between VMS trajectory and migraines and incident CVD and stroke. We hypothesized that migraines would be associated with higher risk of CVD and stroke (ischemic and hemorrhagic combined). We also hypothesized that women with persistent VMS as well as migraines would have higher risk of CVD and stroke compared to women without migraines and with minimal or increasing VMS trajectories. We hypothesized that these associations would persist after adjusting for potential confounders, including gynecological surgery, exogenous estrogen use as well as traditional CVD risk factors.
CARDIA is an ongoing multicenter longitudinal study conducted at four communities (Birmingham, Alabama; Chicago, Illinois; Minneapolis, Minnesota; and Oakland, California) to study CVD risk trends and clinical sequelae from young adulthood. At baseline (Y0, 1985–1986), healthy adults (n=5,115) were recruited from the general population to be balanced on sex, race (White or Black), age (18–24 or 25–30 y) and education (high school or less, or more than high school). Data collection protocols were approved by the Institutional Review Boards (IRBs) of each field center with all participants providing written informed consent. For the Y35 exam, IRB approval for all field centers was performed by a single IRB at the University of Alabama Birmingham.
Details of the study design, recruitment, methodology, and baseline characteristics are described elsewhere.^14^ Of the original cohort of 2787 women, the number of women who attended at least 3 exams between Y15 and Y35 was 1966 (71% of the original cohort). For the purposes of this analysis, we included any woman who responded ≥3 assessments regarding the presence of VMS (n=1966 women) between and including Y15, when questions regarding VMS were first asked, and Y35, the last CARDIA examination conducted. We excluded 12 women who had any CVD event at or before the Y15 exam for a total sample of n=1954 women. Compared to women without these characteristics, women who were excluded were slightly younger (24 years vs. 25 years), more likely to be Black (39% vs. 49%), and more likely to have disadvantageous social determinants of health and health behaviors (less than a high school education, more difficulty paying for basic necessities, higher depressive symptom score, current cigarette use, higher BMI, and higher systolic blood pressure) although prevalence of oral contraceptive pill and other hormonal contraceptive use, other hormone therapy at Y15, perimenopausal symptoms, and any gynecologic surgery was lower among excluded women. Since the prevalence of hormone use was low at Y15, because hormones are used more frequently among women who undergo gynecologic surgery, and also because hormones may not completely resolve VMS, these women were included in the analysis.
The primary outcome of this study was a composite of incident CVD events that included nonfatal coronary heart disease (myocardial infarction [MI], non-MI acute coronary syndrome, and coronary revascularization), stroke (of any type), transient ischemic attack, peripheral arterial disease, hospitalization for heart failure, or death from cardiovascular causes. New cardiovascular and cerebrovascular events were recorded from the baseline examination through 2022. During scheduled study examinations and yearly telephone interviews, participants or their designated proxy were questioned regarding interim hospital admissions, outpatient procedures, and deaths.^15^ For participants who had been hospitalized or who had received an outpatient revascularization procedure, medical records were requested. Vital status was assessed every 6 months; medical and other death records were requested with consent of the next of kin. Over the last 5 years of the study, >90% of the surviving cohort members have been contacted, and follow-up for vital status, assessed using contacts and National Death Index searches, was nearly complete. Two physician members of the Endpoints Surveillance and Adjudication Subcommittee independently reviewed medical records and recorded information to adjudicate possible cardiovascular or cerebrovascular events or underlying causes of death using specific definitions and manual of operations. Disagreements between the 2 physician reviewers were adjudicated by the full committee.^15^ Case definitions for coronary disease, stroke, and heart failure have been previously described;^16^ hemorrhagic and ischemic strokes were combined due to the small number of hemorrhagic strokes.
At every CARDIA examination, standardized protocols were used to collect information on demographics, medical history, lifestyle, behavioral factors, anthropometrics, and cardiovascular risk factors. Beginning at the Y15 examination, women were asked, “Have you experienced hot flashes or night sweats in the past 3 months?” Gynecologic surgery and use of hormonal contraception and other types of hormone therapy were also assessed. Of the 246 women who used hormonal contraception at Y15, approximately 222 used combination estradiol-progestin pills, 7 used progestin-only pills and 8 used medroxyprogesterone acetate injection or levonorgestrel implants, and 9 could not recall formulation type. Of the women who reported any exogenous hormone use at Y15, 38 used oral formulations including an estrogen, 8 used transdermal or injectable formulations, and the remainder could not recall the formulation type or reported non-sex steroid formulations.
At baseline and every 5 years, participants were asked about their medical history. Beginning at Y7, participants were asked “has a doctor or nurse ever said that you have migraine headaches?” Cigarette smoking was assessed by means of an interviewer-administered tobacco questionnaire and classified as current, former or never. Body mass index (BMI) was calculated by dividing measured weight in kilograms by height in meters squared. Blood pressure was measured with participants seated and after 5 minutes of rest. The average of the second and third consecutive measurements was used for analysis. At Y15, serum fasting glucose was measured using hexokinase coupled to glucose 6-phosphate dehydrogenase by Linco Research (St Louis, MO, USA). Lipid measurements using fasting blood samples were collected at baseline and at all follow-up examinations. Total cholesterol and triglyceride (TG) levels were measured enzymatically, high-density lipoprotein cholesterol (HDL-C) was determined after precipitation with dextran sulfate/magnesium chloride, and low-density lipoprotein cholesterol (LDL-C) was calculated by the Friedewald equation.^17^
Using women’s responses to the question “Have you experienced hot flashes or night sweats in the past 3 months?” we used latent class models (fitted by SAS Proc Traj)^18^ to model VMS trajectory, using chronologic age as the x-axis. We analyzed trajectory models with 2 through 10 groups, and model fit was assessed using the Bayesian Information Criterion (BIC) and the Akaike Information Criterion (AIC).^19^ The optimal number of trajectory groups is determined when the BIC and AIC were maximized or when adding more groups did not substantially differentiate between trajectory patterns. We first ran each model with a cubic function; however, the VMS trajectories did not reach a global maximum until we assigned lower polynomial function (quadratic). Next, because trajectory models often find only the local maximum (as opposed to the global maximum) when using default start values, we ran each model using the recommended start parameters and polynomial function for each trajectory group to achieve a model that reached a global maximum (i.e., best-fit polynomial function for each trajectory group within the trajectory model). The posterior predictive probability of group membership was calculated for each model, and participants were assigned to the trajectory group for which they had the greatest posterior predictive probability. The average posterior predictive probabilities were assessed for adequacy of model fit. No missing data were imputed for the present analysis. Women were grouped into 3 minimal VMS over time (n=835), increasing VMS over time (n=521), and persistent VMS over time (n=598) (Figure 1).
Participant characteristics by VMS trajectory group were defined by means, medians, and proportions as appropriate (Table 1), and differences between participants in each trajectory were tested using t-tests, Wilcoxon tests, and χ2 analyses for continuous and categorical characteristics, respectively. We also examined distributions of participants by presence and absence of migraines and trajectory membership. Cox proportional hazards models were used to examine the association between VMS trajectory group and CVD as well as between migraine and CVD after adjusting for age, race, and examination center (Table 3, Model 1). Additional models adjusted for gynecologic history and exogenous hormone use at Y15 (Table 3, Model 2) as well as CVD risk factors at Y15 (Table 3, Model 3) and BMI at Y15 (Table 3, Model 4). Similar models were repeated with stroke as an outcome. Time at risk was calculated as the time from the Y15 examination until the date at which CVD was diagnosed or last contact. The proportional hazards assumption was tested by including a model term for the product of VMS group and migraine and natural log of time at risk; no violations of the assumption were observed. All analyses were conducted using SAS, Version 9.4. A threshold of p<.05 (two-sided) was used to determine statistical significance.
There were 81 incident CVD events including 42 strokes, with a median of 6522 (IQR 6432, 6575) days since the Y15 exam. Characteristics by trajectory group are shown in Table 1. At the Y15 Exam, women with persistent VMS were more likely to be Black, to have less than a high school education, to report histories of migraine headaches and current tobacco use. They were less likely to report oral contraceptive pill use but more likely to report other exogenous hormone therapy use, perimenopausal symptoms, and histories of hysterectomy or oophorectomy. They were also more obese, had slightly higher blood pressure, and had higher triglyceride levels. In unadjusted comparisons (Table 1), women with persistent VMS were more likely to have experienced an incident CVD event or stroke after the Y15 exam compared to women with minimal or increasing VMS. Women with both persistent VMS as well as migraine (n=135) had the poorest CVD risk factor profiles (Table 2).
Table 3 shows the associations between migraines and VMS with incident CVD events. In models that examined the association between migraine with CVD risk, women who had histories of migraines had similar likelihood of experiencing a CVD event as women who did not have histories of migraine after adjustment for other factors (Table 2). Similarly, women had similar CVD risk across minimal, increasing, and persistent VMS trajectories (Table 2). However, women who had histories of both migraines and persistent VMS had higher CVD risk than women who did not have histories of migraine and persistent VMS after adjustment for age, race, examination center, and reproductive factors. Neither the association between hormonal contraceptive use and CVD (OR 1.72, 95% CI 0.69 – 4.31) nor the association between other hormone therapy and CVD (OR 0.68, 95% CI 0.26 – 1.78) was significant. This association was markedly attenuated after adjustment for CVD risk factors. Further adjustment for BMI did not alter associations.
Table 4 shows the association between migraines and VMS with incident strokes. In models that examined the association between migraine with stroke risk, women who had histories of migraines had similar likelihood of experiencing a stroke as women who did not have histories of migraine after adjustment for other factors (Table 4). Similarly, women had similar stroke risk across minimal, increasing, and persistent VMS trajectories (Table 4). However, women who had histories of both migraines and persistent VMS had higher stroke risk than women who did not have histories of migraine and persistent VMS after adjustment for age, race, examination center, and reproductive factors. This association was markedly attenuated after adjustment for CVD risk factors; further adjustment for BMI did not change associations.
We conducted several sensitivity analyses. First, we examined whether associations between migraines/VMS trajectories with outcomes differed by exogenous hormone use by examining whether interactions between these categories and with exogenous estrogen use at Y15 and oral contraceptive use at Y15 were significant at p<0.10. The associations were not significant. Second, we examined whether associations between migraines/VMS trajectories with outcomes differed by potential shared determinants of these factors, including Black race; education level; cigarette use; and BMI by creating separate models that included interaction terms for each of these variables. Interaction terms were not significant at p<0.10 and thus were not included. Third, we examined whether adjustment for family history of CVD significantly altered the pattern of associations; it did not (results not shown). Fourth, we examined whether adjusting for a CVD aggregate risk score^20^ rather than adjustment for individual CVD risk factor levels had a different pattern of associations. In these models, associations between migraine + VMS and CVD remained significant, suggesting that the risk score accounts for less of the variance in associations between migraines + VMS with CVD events (results not shown); this may be due to the fact that the aggregate risk score accounts for diabetes as a dichotomous variable, whereas adjustment for individual CVD risk factor levels accounts for glucose as a continuous variable. Moreover, our adjustment for individual CVD risk factor levels included triglyceride levels, which the risk score does not.
In a population-based prospective cohort, we found that women with histories of migraine who also had persistent VMS had higher risk of CVD and stroke. Although VMS and migraine have both been hypothesized to represent vasoactive conditions that may represent unique CVD risk factors in women, studies have not examined their joint impact upon CVD. History of migraine did not independently increase risk of CVD or stroke, and pattern of VMS did not independently increase risk of CVD or stroke. Although the association between CVD and history of migraine combined with persistent VMS was independent of estrogen use and gynecologic surgery, the association was markedly attenuated by adjustment for CVD risk factors including cigarette use and levels of glucose, lipids, and blood pressure. The results suggest that these two conditions may interact to increase CVD risk, but that risk may potentially be ameliorated through attention to these modifiable risk factors. The results also suggest that history of migraines at 40 years of age combined with presence of VMS may represent early indications of women who may develop abnormal CVD risk factor profiles as they enter middle-age.
Although our report is the first to examine the joint influences of migraine and VMS upon CVD, several reports have examined associations between migraines with CVD, as well as between VMS with CVD. Migraines are particularly common among women in late reproductive age, affecting approximately 17.5% of women.^21^ The majority of migraineurs experience their first attack before the age of fifty years.^22^ Associations between migraine with CVD and between migraine with stroke have long-been recognized in both men and women,^23–25^ although the association may be stronger in women.^26, 27^ Although the precise mechanism is not well-understood, the risk factors may be heterogeneous between women and include vasospasm, hypercoagulability, and inflammation.^28^ Migraine and CVD risk factors may share genetic loci,^29^ with different alleles implicated for migraine with aura; without aura; and both migraine types.^30^ We found that women with migraine histories tended to have less favorable CVD risk factors and this unfavorable risk profile may also be a mechanism for future CVD and stroke, although risk factor levels at the Year 15 examination were generally well-within clinical definitions of normal. Thus, migraines may be an indicator for women at increased risk for future, rather than late reproductive-age, CVD risk factor abnormalities, and the absolute magnitude of increased CVD risk associated with migraine may be low; in one study of approximately forty-thousand female health professionals, the absolute increase in risk associate with migraine was 3.8 additional cases of stroke per year per 10,000 women.^9^ In a Danish cohort, the increased CVD risk associated with migraine was particularly pronounced among persons with aura, in women more than men, and after adjustment for cigarette use and other traditional CVD risk factors.^5^ These studies did not examine interactions with VMS. We may have not found an independent association between migraine and CVD due to the ascertainment of migraine, which enquired after healthcare provider ascertainment of migraine, but which did not enquire after aura, and which therefore may lack precision.
VMS affect almost 80% of women transitioning through menopause, but symptoms vary greatly in severity, frequency, age of onset, and accompanying symptoms.^31^ The etiology of VMS is uncertain, with prior studies suggesting that neither levels of endogenous sex hormones^32^ nor blockage of neurokinin-b receptors completely account for symptomatology.^33^ Previous studies regarding associations between VMS and CVD conflict, perhaps due to differing characterizations of VMS symptoms by severity and/or duration. Our study results are similar to those reported in analysis that pooled individual-level data across 6 prospective studies and noted that there was no association between VMS frequency and CVD risk.^3^ Although greater severity of VMS was associated with higher CVD risk, the analysis did not adjust for CVD risk factors.^3^ Other reports, including another pooled analysis,^4^ have noted that VMS frequency is associated with increased risk of CVD, although these studies noted that much of the association was explained by adjustment for CVD risk factors, suggesting that CVD risk factors are a mediator or confounder for the association.^4^ However, in the Study of Women’s Health Across the Nation (SWAN),^2^ Thurston and colleagues reported that women with VMS on six more days of the week were more likely to experience incident CVD, even after adjustment for CVD risk factors. In another report from SWAN, Tepper and colleagues reported that women could be categorized into several VMS trajectories, including one trajectory that noted persistent severe VMS over multiple visits; these women were also the most likely to have an adverse CVD risk factor profile.^34^ Of note, these studies did not also consider joint effects with migraine. We may have not found independent associations between VMS and CVD due to the smaller number of women and events in CARDIA than in SWAN, or due to our characterization of VMS which relied on persistent VMS over years rather than frequency over the number of days of the week. Other large cohorts have also reported associations between VMS and incident CVD,^35, 36^ although these relied on recall of VMS from years prior and did not examine persistence of VMS over time or other more specific groupings of VMS.
Few prospective studies have examined the association between VMS with stroke. Another systematic review from 2023 noted higher stroke risk among women with VMS only among women < 60 years of age at the time of stroke,^1^ whereas no association was observed for other groups. Although this review was able to stratify by age, it was not able to adjust for CVD risk factor profile, and thus it is possible that associations were mediated through these CVD risk factors. Again, interactions with migraine were not examined. As we examined almost two thousand women, the absolute magnitude of any greater CVD or stroke risk associated with VMS is likely to be small after consideration of CVD risk factors.
Strengths of this analysis include its prospective ascertainment of migraines and VMS in relation to incident events, which reduces the likelihood of recall bias. Another strength was the careful assessment of other CVD risk factors and potential confounders such as exogenous estrogen use and gynecologic surgery. The CARDIA cohort is one of the largest population-based prospective studies of CVD risk factors beginning in early reproductive life, and thus likely to detect strong associations between these risk factors with CVD particularly among women at younger ages. Limitations of the analysis include reliance on self-reported migraine and lack of information regarding severity; this would have biased associations to the null due to misclassification of other types of headaches as migraine and lack of information regarding severity. As mentioned, classification of VMS varies widely between studies and may rely on daily frequency, frequency over years, and severity. Due to the number of strokes, it was not possible to distinguish differences in associations between ischemic vs. hemorrhagic strokes. It is possible that alternate classifications might have yielded a different pattern of associations, although such classifications would require more frequent interaction with participants in order to obtain data that was not biased by recall.
Our findings emphasize the importance of optimization of CVD risk factors in women with persistent VMS and migraines since CVD risk was doubled compared to women without both of these symptoms. Women with persistent VMS as well as migraine history had triple the risk of strokes compared to women without these risk factors, although risk was attenuated with adjustment for CVD risk factors. Our results should also be reassuring to women with either migraine history or persistent VMS over time, whose risk for CVD events in late middle-life does not appear to be significantly elevated. Future studies are needed to examine why women with both migraines and VMS have the poorest CVD risk factor profiles and whether these histories can aid in risk stratification. Studies are also needed to examine whether this increased risk is limited to women who experience migraine with aura, and whether therapies for migraines as well as non-hormone therapies for VMS are associated with future CVD risk.