|Home | About | Journals | Submit | Contact Us | Français|
Emerging evidence suggests that women with menopausal vasomotor symptoms (VMS) have increased cardiovascular disease (CVD) risk as measured by surrogate markers. We investigated the relationships between VMS and clinical CVD events and all-cause mortality in the Women's Health Initiative Observational Study (WHI-OS).
We compared the risk of incident CVD events and all-cause mortality between four groups of women (total N=60,027): (1) No VMS at menopause onset and no VMS at WHI-OS enrollment (no VMS [referent group]); (2) VMS at menopause onset, but not at WHI-OS enrollment (early VMS); (3) VMS at both menopause onset and WHI-OS enrollment (persistent VMS [early and late]); and (4) VMS at WHI-OS enrollment, but not at menopause onset (late VMS).
For women with early VMS (N=24,753), compared to no VMS (N=18,799), hazard ratios (HRs) and 95% confidence intervals (CIs) in fully-adjusted models were: major CHD, 0.94 (0.84, 1.06); stroke, 0.83 (0.72, 0.96); total CVD, 0.89 (0.81, 0.97); and all-cause mortality, 0.92 (0.85, 0.99). For women with persistent VMS (N=15,084), there was no significant association with clinical events. For women with late VMS (N=1,391) compared to no VMS, HRs and 95% CIs were: major CHD, 1.32 (1.01, 1.71); stroke, 1.14 (0.82, 1.59); total CVD, 1.23 (1.00, 1.52); and all-cause mortality, 1.29 (1.08, 1.54).
Early VMS were not associated with increased CVD risk. Rather, early VMS were associated with decreased risk of stroke, total CVD events, and all-cause mortality. Late VMS were associated with increased CHD risk and all-cause mortality. The predictive value of VMS for clinical CVD events may vary with onset of VMS at different stages of menopause. Further research examining the mechanisms underlying these associations is needed. Future studies will also be necessary to investigate whether VMS that develop for the first time in the later postmenopausal years represent a pathophysiologic process distinct from classical perimenopausal VMS.
Cardiovascular disease (CVD) is the leading cause of mortality in women,1 and the risk of CVD in women increases dramatically after menopause.2 The majority of menopausal women experience vasomotor symptoms (VMS).3 Higher blood pressure,4–6 cholesterol levels,6 and body mass index,6 as well as increased risk of subclinical CVD7 have been reported among women with versus without hot flashes. Furthermore, recent analyses from two major clinical trials, the Women's Health Initiative Hormone Therapy Clinical Trials8 and the Heart and Estrogen/progestin Replacement Study (HERS),9 reported that the elevated risk of coronary heart disease (CHD) with late hormone therapy (HT) use was concentrated among those with VMS.
It is important to gain a better understanding of the possible underlying differences in CVD risk between women with and without VMS, given that most menopausal women have VMS and that many women experience these symptoms well into the later postmenopausal years,10–12 a time when CVD risk accelerates. To our knowledge, the independent contributions of VMS in predicting incident clinical CVD events have not been reported, and no previous study has distinguished between VMS at the onset of menopause and VMS occurring later in menopause. VMS were more common among women participating in the WHI Observational Study (WHI-OS), compared to the WHI HT Clinical Trials, as the WHI HT Clinical Trials excluded women with significant VMS while the WHI-OS was open to symptomatic women excluded from the WHI HT Clinical Trials. Thus, the WHI-OS cohort provides an opportunity to examine these relationships in a population enriched with women who experienced VMS.
The objective of our study was to investigate whether VMS predict incident CVD events and all-cause mortality in the WHI-OS and whether the relationship is different for VMS at menopause onset and VMS later in menopause.
As previously described,13, 14 the WHI is comprised of the WHI Clinical Trials and the WHI-OS. The WHI-OS is an ongoing longitudinal study examining the relationships between new risk indicators for disease and health outcomes in a large prospective cohort of postmenopausal women. The WHI-OS enrolled 93,676 women between 1994 and 1998 at 40 U.S. clinical centers. Study methods of the WHI-OS have been described in detail previously.15 Briefly, the WHI-OS included women who were either ineligible or unwilling to participate in the WHI Clinical Trials, or were recruited through a direct invitation to participate in the OS. Women were included if they were aged 50–79 years at the time of enrollment, were postmenopausal, were able and willing to provide written informed consent, and were unlikely to move from their area of residence for at least 3 years.16 Postmenopausal status was defined as amenorrhea for at least 6 months if aged ≥55 years and amenorrhea for at least 12 months if aged 50–54 years. Women were excluded if they had any medical condition with predicted survival of less than 3 years, had conditions which might impede adherence to the protocol or retention in the study, or were participating in either the WHI Clinical Trials or another randomized intervention trial. Signed informed consent was obtained from all participants, and the study was reviewed and approved by human subjects review committees at each participating institution.
All WHI-OS participants completed questionnaires at baseline that asked about demographic and lifestyle variables, as previously described.13 Anthropometric measurements and blood pressure were obtained at the initial screening visit. Body mass index was calculated as weight (in kg) divided by height (in m2). Questionnaires were used to collect information about age, ethnicity, age at menopause, smoking status, history of hysterectomy or bilateral oophorectomy, educational level, aspirin use, statin use, and physical activity (reported in this analysis as total metabolic equivalents [MET] hours/week, classified as 0, < 5, 5 – <12, or ≥12 MET hours/week). Participants were asked whether they had ever been told by a physician that they had hypertension or high blood pressure, diabetes or high blood glucose when they were not pregnant, high cholesterol requiring pills, or a family history of myocardial infarction at a young age in first-degree relatives (<55 years for male relatives and <65 years for female relatives). In addition, participants completed a questionnaire about HT use, and participants were classified as never, past, or current users of HT. Current or past HT use was defined as use of an estrogen- or progesterone- containing pill or transdermal patch for at least 3 months following menopause. Vaginal HT was not included in current or past HT use. Age at menopause was defined as previously described.8 A self-administered questionnaire was used to ask women about the ages at which they first and last had VMS, defined in this manuscript as hot flashes or night sweats. Women were considered to have VMS at menopause onset if age at first hot flash or night sweats was less than or equal to age at menopause. Women were also asked at study enrollment about the presence of VMS during the preceding four weeks. If VMS were present, participants were asked to rate the symptoms as mild (symptom did not interfere with usual activities), moderate (symptom interfered somewhat with usual activities), or severe (symptom was so bothersome that usual activities could not be performed). For this analysis, we chose a priori to categorize VMS at enrollment dichotomously as present (mild, moderate or severe) or absent, in order to mirror the dichotomous data available for VMS at menopause onset.
The clinical outcomes evaluated in this analysis were major coronary heart disease (CHD), stroke, total CVD events (major CHD and stroke), and all-cause mortality. CHD was defined as acute MI requiring overnight hospitalization, silent MI determined from serial electrocardiograms, or CHD death, as previously described.17 Stroke diagnosis was based on rapid onset of a neurologic deficit lasting more than 24 hours, supported by imaging studies when available. Non-fatal events in the WHI-OS were locally adjudicated by trained physician reviewers. A sample of the locally-adjudicated events was reviewed by central cardiovascular adjudicators.18 All deaths were centrally adjudicated. Locally-adjudicated myocardial infarctions and strokes were confirmed on central review for 87%18 and 91% (Women's Health Initiative Clinical Trial and Observational Study Semi-Annual Progress Report, March 1, 2004 to August 31, 2004) of the cases, respectively.
This analysis includes women who had no prior history of CVD or cancer at baseline and who had data reported on VMS status at both menopause onset and WHI-OS enrollment. Of the 93,676 postmenopausal women enrolled into the WHI-OS, 78,249 had no prior history of CVD or cancer. 77,631 (99.2%) of these women reported information on VMS status at enrollment, and 60,773 (77.7%) of these women reported information on VMS status at menopause onset. Our study sample included 60,027 women who fulfilled all inclusion criteria. As of August 2007, median follow-up duration for the study population was 9.7 years (interquartile range, 8.7–10.9); 4.3% withdrew or were lost to follow-up and 6.7% died. The annual follow-up response rate was >98% for those scheduled for a follow-up contact.
Categorical baseline characteristics were standardized by age using the direct method. The standard population was the age distribution of the whole cohort. Women were classified into four groups according to VMS status, as follows: (1) No VMS at menopause onset and no VMS at WHI-OS enrollment (hereafter referred to as no VMS [referent group]); (2) VMS at menopause onset, but not at WHI-OS enrollment (hereafter referred to as early VMS); (3) VMS at both menopause onset and at WHI-OS enrollment (hereafter referred to as persistent VMS [early and late]); and (4) VMS at WHI-OS enrollment, but not at menopause onset (hereafter referred to as late VMS). Cox proportional hazards models were used to assess the independent contribution of VMS in predicting incident CVD events and all-cause mortality. The assumption of proportional hazards was tested by examining plots of the baseline hazard by VMS and by testing interaction terms of VMS by time; the assumption was valid. The time to event was calculated from the date of enrollment and the date of the outcome. The discrete logistic model was used for handling ties. Minimally-, partially-, and fully-adjusted models were fit for outcomes stratified by 10-year age group. Minimally-adjusted models were adjusted for age and ethnicity. Partially-adjusted models were additionally adjusted for education, body mass index, smoking, hysterectomy status, bilateral oophorectomy status, aspirin use, statin use, hormone therapy status at baseline, and physical activity. Fully-adjusted models were additionally adjusted for hypertension, hypercholesterolemia, diabetes, and family history of CHD. In order to test for differential associations between VMS status and CVD events as a function of HT use, an interaction term of VMS status and HT use was included. Analyses were also performed separately according to HT use (never, past, and current) and according to whether or not women had a history of surgical menopause (defined as history of bilateral oophorectomy).
Comparing incident total CVD events between women with versus without VMS at menopause onset, our study had >99% power (with a 2-sided alpha level of 0.05) to detect a HR of 1.2 and higher. All P values were two-tailed, and P values <0.05 were considered statistically significant. All analyses were performed using SAS software, version 9.1 (SAS Institute, Cary, NC).
Mean age of the women included in this analysis was 63.3 years (range, 50 to 79 years), and women were a mean of 14.4 years since menopause (range, 0 to 54 years) at the time of WHI-OS enrollment. Table 1 shows the baseline characteristics of the study population by VMS status, standardized by age. Most women had either no VMS or early VMS.
The associations between VMS status and incident CVD events and all-cause mortality are shown in Table 2. Overall, there was no significant association between VMS and the four clinical outcomes examined (data not shown). For women with early VMS compared to no VMS, there was a decreased risk of stroke, total CVD events, and all-cause mortality in minimally-adjusted, partially-adjusted, and fully-adjusted models (Table 2). For women with persistent VMS (early and late), there was no significant increase or decrease in risk of major CHD, stroke, total CVD, or all-cause mortality compared to women with no VMS in partially- and fully-adjusted models (Table 2). In contrast, for women with late VMS compared to no VMS, there was an increased risk of major CHD events, stroke, total CVD events, and all-cause mortality in the minimally-adjusted models (Table 2). In the partially-adjusted models, there was an increased risk of major CHD events, total CVD events, and all-cause mortality in women with late VMS compared to no VMS. In the fully-adjusted models, the significantly increased risk of major CHD and all-cause mortality among women with late VMS persisted.
Analyses performed after excluding women with a history of surgical menopause were similar to those in the overall cohort, though some of the associations were not statistically significant (results for total CVD and all-cause mortality are shown in the Figure). In analyses performed among the smaller group of women who did have a history of surgical menopause (N=9,804), most of the associations were not statistically significant, except for decreased risks of stroke and total CVD for women with early VMS (HRs [95% CIs]: 0.67 [0.48, 0.93] and 0.76 [0.61, 0.94], respectively). Analyses performed separately according to HT use provided results similar to those in the overall cohort, though most of the associations did not remain statistically significant; results for total CVD and all-cause mortality among women who never used HT are shown in the Figure. Formal testing using interaction terms showed no evidence of an interaction between HT use and VMS at either menopause onset or at WHI-OS enrollment in predicting CVD events or all-cause mortality. .
In summary, we found that a history of early VMS was not associated with increased risk of future clinical CVD events or all-cause mortality. Indeed, the risks of stroke, total CVD events, and all-cause mortality were reduced among women with early VMS. In contrast to these inverse associations, we found that late VMS were associated with increased risk of major CHD and all-cause mortality, even after adjustment for traditional CVD risk factors.
A growing body of literature has examined the possible links between menopausal VMS and surrogate CVD risk markers. Higher ambulatory blood pressure has been reported among women with versus without hot flashes.4, 5 Reduced heart rate variability, a marker of increased cardiovascular risk, has been reported during hot flashes.19 A cross-sectional analysis of the Eindhoven Perimenopausal Osteoporosis Study reported that women with versus without VMS had higher blood pressure, cholesterol levels, and body mass index.6 Analyses of perimenopausal and postmenopausal women enrolled in the Study of Women's Health Across the Nation (SWAN) Heart Study reported that hot flashes were associated with increased risk of subclinical CVD, including impaired endothelial function,7 increased aortic calcification,7 and increased carotid intima media thickness.20
Our findings are not at odds with previous studies reporting an association between VMS and adverse cardiovascular risk as measured by surrogate markers. Rather, our data confirm that women who reported VMS (particularly women with late VMS) had higher prevalence of several CVD risk factors, in agreement with these previous studies. We, however, found that a history of early VMS did not translate into an increased risk of clinical CVD events. These findings highlight the importance of outcome measures which focus on clinical events rather than surrogate markers. Menopause is a normal phase of women's lives, and most women experience VMS over the course of the menopausal transition.3 Our findings that early VMS were not associated with increased risk of clinical CVD events or all-cause mortality are therefore potentially relevant to a large number of women. It is interesting that one recent study reported that a history of both hot flashes and night sweats (but not hot flashes alone) was associated with a decreased risk of death among older postmenopausal women;21 this study did not differentiate between early and late VMS and did not examine CVD events. The mechanisms underlying the inverse associations we report between early VMS and CVD events are not currently known. One possibility is that perimenopausal VMS represent a physiologic response to the normal perimenopausal hormonal fluctuations, and the absence of these symptoms may signify a blunted vascular response to these hormonal changes. In support of this hypothesis, a recent study reported that recently menopausal women with severe versus no hot flashes exhibited greater nitroglycerin-induced (endothelium-independent) vasodilatation.22 Other studies, however, reported impaired flow-mediated (endothelium-dependent) dilation among women with any7 or clinically-significant23 hot flashes. Together these studies suggest that the potential vascular mechanisms linking early VMS to decreased CVD risk are not mediated through the endothelium. Further studies examining the mechanisms underlying these associations are needed.
The women who reported late VMS represent a small subset of our study population, and the associations we report between late VMS and increased risk of major CHD and all-cause mortality should therefore be interpreted as preliminary and hypothesis-generating. Adjustment for traditional coronary risk factors in the fully-adjusted models attenuated these associations, suggesting that traditional coronary risk factors may be in the causal pathway linking late VMS to increased risk of major CHD and all-cause mortality. The fact that these relationships remained significant after adjustment for traditional coronary risk factors, however, suggests that other currently unidentified mechanisms also increase CVD risk among women with late VMS. Our adjustment for HT use makes it unlikely that recently-initiated HT explains the increased risk of CHD and all-cause mortality reported among women with late VMS. Future studies are necessary to investigate the currently unknown pathophysiologic mechanisms underlying the development of VMS many years after menopause. These symptoms may be a marker of vascular instability or an early manifestation of cardiac ischemia. Of interest, recent analyses of the WHI HT Clinical Trials8 and HERS9 reported that the increased CHD risk associated with late HT use was amplified among the subset of women who reported VMS at study enrollment.8,9 In addition, another recent study reported that a longer duration of reported hot flashes was associated with increased aortic calcification among HT users.24 Since many women experience hot flashes in the later postmenopausal years,10–12 future studies which prospectively evaluate the complex inter-relationships between VMS and CVD events at different menopausal stages are warranted. In addition, future studies which objectively characterize the potential physiologic differences between VMS occurring in early versus late postmenopausal women are needed.
Strengths of our study include the geographic, ethnic, and socioeconomic diversity of the WHI population and the community-based study population, which enhances the generalizability of our findings. In addition, our study is the first study to examine the independent contributions of VMS in predicting incident clinical CVD events.
A limitation of our study is the reliance on retrospective self-report of VMS. While physiologic measurement of hot flushes via sternal skin conductance is possible, this technique is not practical for a study large enough to evaluate clinical outcomes. Furthermore, subjective (not objective) assessment of hot flushes is employed in clinical practice, though ideally these symptoms would be elicited in a real-time setting rather than a historical one. The observational nature of our study (a limitation inherent to all studies examining VMS as a clinical predictor) precludes conclusions about causality. It is possible that residual confounding may contribute to the associations we report. An important limitation is our inability to disentangle the complex relationships between VMS and HT use in an observational setting. We acknowledge that past or current HT use may have altered VMS status classification in some women. Additionally, the decision to initiate HT is often (but not solely) influenced by the presence of VMS, and the subjective experience of VMS is linked to HT use,25 which is recognized to influence CVD risk independently.17 We adjusted for HT use in the multivariable models, and formal testing using interaction terms showed no evidence of an interaction between HT use and VMS at either menopause onset or at WHI-OS enrollment in predicting CVD events or all-cause mortality. While most of the associations we report for the entire study population did not remain statistically significant in analyses restricted to women who never used HT, it is not clear whether this indicates the presence of residual confounding by HT or inadequate power to detect these relationships in the smaller subsets.
We found no evidence of an increased risk of clinical CVD events or all-cause mortality among women with early VMS. Rather, early VMS were associated with decreased risk of stroke, total CVD events, and all-cause mortality. We did observe an increased risk of major CHD and all-cause mortality among the small subset of women with late-onset VMS. Our results suggest that a history of perimenopausal VMS is not likely to be an independent CVD risk factor and may even be inversely associated with CVD risk. The predictive value of VMS for clinical CVD events may vary with onset of VMS at different stages of menopause. Future studies will be necessary to investigate whether VMS that develop for the first time in the later postmenopausal years represent a pathophysiologic process distinct from classical perimenopausal VMS.
In the Women's Health Initiative Observational Study, perimenopausal vasomotor symptoms were associated with decreased risks of stroke, total CVD events, and all-cause mortality. In contrast, late vasomotor symptoms were associated with increased CHD risk and all-cause mortality. The predictive value of VMS for clinical CVD events may vary with onset of VMS at different stages of menopause.
WHI investigators Program Office: (National Heart, Lung, and Blood Institute, Bethesda, Maryland) Elizabeth Nabel, Jacques Rossouw, Shari Ludlam, Joan McGowan, Leslie Ford, and Nancy Geller.
Clinical Coordinating Center: (Fred Hutchinson Cancer Research Center, Seattle, WA)
Ross Prentice, Garnet Anderson, Andrea LaCroix, Charles L. Kooperberg, Ruth E. Patterson, Anne McTiernan; (Medical Research Labs, Highland Heights, KY) Evan Stein; (University of California at San Francisco, San Francisco, CA) Steven Cummings.
Clinical Centers: (Albert Einstein College of Medicine, Bronx, NY) Sylvia Wassertheil-Smoller; (Baylor College of Medicine, Houston, TX) Aleksandar Rajkovic; (Brigham and Women's Hospital, Harvard Medical School, Boston, MA) JoAnn E. Manson; (Brown University, Providence, RI) Charles B. Eaton; (Emory University, Atlanta, GA) Lawrence Phillips; (Fred Hutchinson Cancer Research Center, Seattle, WA) Shirley Beresford; (George Washington University Medical Center, Washington, DC) Lisa Martin; (Los Angeles Biomedical Research Institute at Harbor-UCLA Medical Center, Torrance, CA) Rowan Chlebowski; (Kaiser Permanente Center for Health Research, Portland, OR) Yvonne Michael; (Kaiser Permanente Division of Research, Oakland, CA) Bette Caan; (Medical College of Wisconsin, Milwaukee, WI) Jane Morley Kotchen; (MedStar Research Institute/Howard University, Washington, DC) Barbara V. Howard; (Northwestern University, Chicago/Evanston, IL) Linda Van Horn; (Rush Medical Center, Chicago, IL) Henry Black; (Stanford Prevention Research Center, Stanford, CA) Marcia L. Stefanick; (State University of New York at Stony Brook, Stony Brook, NY) Dorothy Lane; (The Ohio State University, Columbus, OH) Rebecca Jackson; (University of Alabama at Birmingham, Birmingham, AL) Cora E. Lewis; (University of Arizona, Tucson/Phoenix, AZ) Cynthia A Thomson; (University at Buffalo, Buffalo, NY) Jean Wactawski-Wende; (University of California at Davis, Sacramento, CA) John Robbins; (University of California at Irvine, CA) F. Allan Hubbell; (University of California at Los Angeles, Los Angeles, CA) Lauren Nathan; (University of California at San Diego, LaJolla/Chula Vista, CA) Robert D. Langer; (University of Cincinnati, Cincinnati, OH) Margery Gass; (University of Florida, Gainesville/Jacksonville, FL) Marian Limacher; (University of Hawaii, Honolulu, HI) J. David Curb; (University of Iowa, Iowa City/Davenport, IA) Robert Wallace; (University of Massachusetts/Fallon Clinic, Worcester, MA) Judith Ockene; (University of Medicine and Dentistry of New Jersey, Newark, NJ) Norman Lasser; (University of Miami, Miami, FL) Mary Jo O'Sullivan; (University of Minnesota, Minneapolis, MN) Karen Margolis; (University of Nevada, Reno, NV) Robert Brunner; (University of North Carolina, Chapel Hill, NC) Gerardo Heiss; (University of Pittsburgh, Pittsburgh, PA) Lewis Kuller; (University of Tennessee Health Science Center, Memphis, TN) Karen C. Johnson; (University of Texas Health Science Center, San Antonio, TX) Robert Brzyski; (University of Wisconsin, Madison, WI) Gloria E. Sarto; (Wake Forest University School of Medicine, Winston-Salem, NC) Mara Vitolins; (Wayne State University School of Medicine/Hutzel Hospital, Detroit, MI) Michael Simon.
Women's Health Initiative Memory Study: (Wake Forest University School of Medicine, Winston-Salem, NC) Sally Shumaker.
Funding/support: The WHI program is funded by the National Heart, Lung, and Blood Institute, National Institutes of Health, U.S. Department of Health and Human Services through contracts N01WH22110, 24152, 32100-2, 32105-6, 32108-9, 32111-13, 32115, 32118-32119, 32122, 42107-26, 42129-32, and 44221.
Financial disclosure/conflicts of interest Dr. Szmuilowicz reported that she was previously a sub-investigator in a clinical trial of a diabetes treatment (GLP-1 agonist) sponsored by Sanofi-aventis, and she received no financial compensation. Dr. Seely reported that she received a Bayer Health Care investigator-initiated grant. Dr. Howard reports: consultant for Merck/Schering-Plough; research support from donation of drugs from Merck/Shering-Plough; lectures for Merck/Schering-Plough. None of the other authors reported financial disclosures.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.