| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Departments of Epidemiology and Biostatistics, University of Michigan School of Public Health (M.R.S., M.J., J.F.R., D.M., R.L.), Ann Arbor, Michigan 48104; University of Pittsburgh (K.A.M., K.S.-T.), Pittsburgh, Pennsylvania 15260; University of California (B.L.), Davis, California 95616; and Merck & Co., Inc. (R.P.), Rahway, New Jersey 08889
Address all correspondence and requests for reprints to: Dr. MaryFran R. Sowers, Department of Epidemiology, University of Michigan School of Public Health, 339 East Liberty Street, Suite 310, Ann Arbor, Michigan 48104. E-mail: mfsowers{at}umich.edu.
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Methods: Longitudinally, we studied 3302 participants from the Study of Women’s Health Across the Nation, aged 42–52 yr at baseline and self-identified as African-American (28%), Caucasian (47%), Chinese (8%), Hispanic (8%), or Japanese (9%). Serum samples from baseline and years 2001, 2003, and 2005 were assayed for estradiol and FSH. Hormone concentrations were related to CVD markers, including fibrinogen, factor VII-c, plasminogen activator inhibitor-1 (PAI-1), tissue plasminogen activator, and human serum C-reactive protein (hsCRP).
Results: Lower estradiol levels were associated with higher levels of PAI-1 and tissue plasminogen activator, but there were no significant relationships with fibrinogen, factor VII-c, or hsCRP. Higher FSH concentrations were associated with higher PAI-1 and factor VII levels, but lower fibrinogen and hsCRP levels. Transitions from premenopause and early perimenopause to postmenopause were not associated with significant differences in levels of hemostatic factors. The hsCRP concentrations were approximately 25% higher, and the PAI-1 concentrations approximately 20% lower among women who initiated hormone therapy, compared with nonusers.
Summary: Endogenous estrogens may reduce CVD risk via modulation of fibrinolytic factors, but not coagulation or inflammatory markers. Notably, conclusions derived from studies of exogenous hormones and CVD risk may not parallel or explain the effects of endogenous hormones or perimenopausal hormone changes on CVD risk.
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Studies of exogenous hormone use in postmenopausal women showed that estrogen replacement therapy lowered plasminogen activator inhibitor-1 (PAI-1) plasma levels, leading to speculation that a cardioprotective effect of estrogen replacement therapy could be expressed through maintenance of a more favorable fibrinolytic balance (7). Two reports indicated that premenopausal women had lower PAI-1 concentrations than postmenopausal women (8, 9), with the inference that differences in PAI-1 concentrations were related to the difference in estrogen concentrations in pre- vs. postmenopausal women. In the baseline examination of the Study of Women’s Health Across the Nation (SWAN), a prospective study of 3302 women transitioning the menopause, the free estradiol index was weakly, but significantly, correlated with PAI-1 concentrations (10).
We related hemostatic and inflammation factors to the endogenous hormones, estradiol (E2) and serum FSH concentrations observed over a 5-yr period and to menopause status in women transitioning the menopause. The hemostatic markers included the fibrinolytic factors PAI-1 and tissue plasminogen activator (tPA), the coagulation factors fibrinogen and factor VII-c, and the acute phase protein, human serum C-reactive protein (hsCRP). We hypothesized that lower E2, higher FSH, and transition to postmenopause would be associated with higher levels of fibrinolytic factors and hsCRP.
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
SWAN is a prospective, multiethnic, multidisciplinary study of the natural history of the menopausal transition that is being conducted in community-based groups of women located in Boston; Chicago; the Detroit area; Los Angeles; Hudson County, NJ; Pittsburgh; and Oakland, CA. Recruitment was a two-stage process commencing with a 15-min cross-sectional survey among 16,065 women, aged 40–55 yr and living in the geographic area defined by the clinic sites. This initial stage served as the sampling frame for the second recruitment stage, which led to the enrollment of 3302 menstruating women, aged 42–52 yr. Enrolled women were not using exogenous hormone preparations that could affect ovarian function in the 3 months before enrollment, had at least one menstrual period in the 3 months before enrollment, and self-identified with the site’s designated race/ethnic group. These criteria precluded enrollment of postmenopausal women or premenopausal women using oral contraceptives or hormone replacement to the longitudinal cohort.
Women who were Caucasian or a member of a designated race/ethnic group were enrolled at each site, including African-American women at Boston, Chicago, the Detroit area, and Pittsburgh as well as Japanese, Chinese, and Hispanic women at Los Angeles, Oakland, and Hudson County, NJ, respectively. Additional information has been published about the eligibility criteria, sampling frames, and participant characteristics (11). Data were collected via protocols reviewed and endorsed by an appropriate institutional review board at each site. This report is based on data from the four annual examinations at which hemostatic factors were measured (baseline and follow-up year 2001, 2003, and 2005 examinations).
Measures
Menopausal status was based on self-report of decreased predictability in the time between menses in the previous 3 months (early perimenopausal), no decreased predictability in the same time period (premenopausal), no menses for 3–11 months (late perimenopausal), or no menses for 12 or more months (menopausal). Height (centimeters) and weight (kilograms) were measured with stadiometers and calibrated scales and were used to calculate body mass index [BMI; weight (kilograms)/height (meters)2]. Waist was measured at the narrowest part of the torso (centimeters).
Assays
At baseline, blood was drawn during d 2–5 of the follicular phase of the menstrual cycle and after fasting. In follow-up examinations, blood draws became increasingly less like to occur in the d 2–5 follicular phase window because menstrual bleeding became increasingly unpredictable in women approaching the last menstrual period (64% at follow-up year 2001, 48% at yr 2003, and 27% at yr 2005). Samples for hormone assays were kept at room temperature for 30–60 min and then refrigerated for 30–60 min. Other samples were refrigerated for up to 2 h, spun, separated, frozen at –20 C (or lower), and sent on dry ice to either the CLIA-certified CLASS laboratory at the University of Michigan (for assay of E2, testosterone, SHBG, FSH, or TSH) or the Medical Research Laboratories (for assay of fibrinogen, hsCRP, tPA, PAI-1, and factor VII-c). Throughout the study, the MRL Laboratory participated in the certification by the National Heart, Lung, and Blood Institute (12).
tPA was measured in plasma using a double antibody in an ELISA (IMUBIND tPA ELISA, American Diagnostica, Greenwich, CT). The assay uses human single-chain tPA as a standard calibrated against an international standard (National Institute for Biological Standards and Control, Hertfordshire, UK). Monthly interassay coefficients of variation (CVs) were 4.7–8.7% and 3.8–7.8% at mean concentrations of 5.6 and 11 ng/dl, respectively.
Plasma PAI-1 was measured with a sandwich procedure using a solid phase monoclonal antibody and an enzyme-labeled goat second antiserum (IMUBIND plasma PAI-1 ELISA, American Diagnostica). The monthly interassay CVs were 5–9% and 4–9% at mean concentrations of 7 and 22.5 ng/dl, respectively.
Fibrinogen was measured in frozen citrated plasma on an MLA ELECTRA 1400C (Medical Laboratory Automation, Inc., Mt. Vernon, NY) using a clot-based turbidometric detection system. The monthly interassay CVs were 2.3–3.5% and 2.6–3.6% at mean concentrations of 250 and 140 mg/dl, respectively.
Factor VII-c activity was measured in frozen citrated plasma on the MLA ELECTRA 1400C (Medical Laboratory Instrumentation, Inc.) using a turbidometric detection system and factor VII-deficient plasma (George King Bio-Medical, Overland Park, KS) in preparation of the standard curve. The monthly interassay CVs were approximately 7.8%, 5%, and 4% for mean activities of 8%, 45%, and 99%, respectively.
hsCRP was quantified using an ultrasensitive rate immunonephelometric method (hsCRP on BN 100, Dade-Behring, Marburg, Germany). The method is based on monitoring light scattering during agglutination of CRP to polystyrene particles coated with monoclonal antibodies to CRP. The sensitivity of the assay (lowest detectable concentration) was 0.03 mg/dl. The CVs at CRP concentrations of 0.05 and 2.2 mg/dl were 10–12% and 5–7%, respectively.
Serum FSH concentrations were measured with a two-site chemiluminometric immunoassay with CVs of 12.0% and 6.0%. SHBG was a de novo two-site chemiluminescent assay with CVs of 9.9% and 6.1%. Serum E2 concentrations were measured with a modified, off-line ACS:180 (E2-6) immunoassay with CVs of 10.6% and 6.4%. Total E2 was indexed to SHBG (free estradiol index = 100 x total estradiol/272.11 x SHBG) to estimate nonbound E2 activity.
Statistical methods
Variables for fasting status, time of day of blood draw, and day of blood draw (d 2–5 of the early follicular phase) were eligible for inclusion in models because tPA, PAI-1, and E2 had significant diurnal variation, and tPA, PAI-1, E2, and BMI values were higher in women whose blood was drawn at times other than d 2–5 of the menstrual cycle. A variable for site and ethnicity was included in all models to account for sampling design.
Data from women treated with anticoagulants (n = 17) were excluded from analyses for the particular examination in which women reported using anticoagulants. Data from women who reported using aspirin were retained, because the reason for aspirin use could not be explicitly discerned, and there was no difference in the hemostatic factor values among women who reported aspirin use compared with those who did not. When hormone therapy (HT) use was reported at postbaseline examinations, the data were segregated into an HT stratum. Data from women with hysterectomy were censored at the time of its report.
Continuous variables, other than age, were log transformed to satisfy model assumptions including normally distributed residuals. Consistency of associations was evaluated including and excluding those women with TSH values outside the euthyroid range of 0.5–5.0 mIU/ml.
Data were first described with cross-sectional analyses and are reported in Tables 1
and 2. Subsequently, longitudinal analyses were used to describe the associations of hormones and menopause status variables with hemostatic factor concentrations using linear mixed models (SAS Proc Mixed, SAS Institute, Inc., Cary, NC) and accounting for the autocorrelation of repeated measures (Tables 3 and 4). The longitudinal linear mixed models included age (as a time varying covariate) and menopause status or reproductive hormones in relation to the hemostatic factors. These variables were treated as main effects (or combined with age as an interaction term) to describe change with time. Models were fit separately for each of the hemostatic factors. Waist circumference and BMI were evaluated in models as time-varying covariates. Self-reported race/ethnicity, site, and baseline smoking behavior were entered into models as single-time covariates. Data from longitudinal models were incorporated into bar graphs to facilitate the interpretation of the ß coefficients for menopause status (Fig. 1) and a range of representative values of FSH or E2 (Figs. 2 and 3); additional effects of hormone x age interaction, when statistically significant, were presented, with age held constant at 50 yr.
|
|
|
|
|
|
|
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Table 1
shows the unadjusted cross-sectional mean values of the hemostatic factors and hormones (or their indices) at each of the four examinations at which CVD markers were assayed and according to the presence or absence of HT.
Factor VII-c levels were significantly lower in African-American women compared with Caucasian, Chinese, Hispanic, and Japanese women (data not shown).
Stages of menopausal status
At baseline, 54% of women were classified as being premenopausal, and the remaining 46% were classified as being in the early perimenopause (Table 2). Only 5% of women remained classified as premenopausal at the fifth follow-up examination. As a condition of enrollment in the longitudinal cohort, women did not use HT at the baseline examination; however, by the fifth follow-up examination, 20% of women, cumulatively, had used HT.
When evaluated with longitudinal models, there were no statistically significant associations of hemostatic factor levels with menopause status as women transitioned from premenopause to the early postmenopause (Table 3 and Fig. 1). All data were adjusted for BMI, waist circumference. and smoking, the covariates that were independently associated with the hemostatic factors.
Reproductive hormones and hemostatic factors
Endogenous E2 concentrations were negatively associated with the fibrinolytic markers, PAI-1 and tPA, and the association with PAI-1 became more positive over time as E2 concentrations declined (P < 0.0001; Table 4 and Fig. 2). There was no association of E2 with hsCRP, fibrinogen, or factor VII-c.
FSH concentrations were positively associated with PAI-1 and factor VII-c (Table 4). FSH was negatively associated with fibrinogen and hsCRP concentrations (Fig. 3).
HT use
HT use, initiated after the baseline examination, was associated with significant differences in hemostatic factors (Fig. 1). The degree of impact was typically greater with longer HT use, as estimated with the significant age x HT interaction term (Table 3). Notably, hsCRP concentrations were 25% higher among the HT users compared with those in premenopausal and early perimenopausal women. Fibrinogen concentrations were 3% lower. PAI-1 concentrations were 21% lower in women with HT use and declined 4% with each additional year of use. On the average, tPA concentrations were approximately 6% lower in women using HT and declined 3.5%/yr of HT use. Factor VII-c was increased among women using HT by 5% compared with levels in pre- and perimenopausal women and continued a 1% increase with each additional year of use.
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
This is, to our knowledge, the first published longitudinal study of endogenous E2 concentrations in relation to the fibrinolytic markers, especially PAI-1, and the findings are consistent with the limited literature data. Cell culture systems have been used to demonstrate that 17ß-estradiol inhibits the synthesis of PAI-1 in endothelial cells (14), and there have been reports of potential estrogen (and glucocorticoid-progesterone) response elements in genes coding for tPA and PAI-1 (15). The lower PAI-1 levels associated with subsequent HT use in this study are consistent with the lower PAI-1 concentrations previously reported (16, 17, 18, 19, 20). Scarabin et al. (21) reported significantly lower PAI-1 levels with oral, but not transdermal, E2 replacement therapy in healthy postmenopausal women. These lower levels may be related to increased clearance, an interpretation provided by studies of oral contraceptive preparations in which ethinyl E2 administration led to reductions in urinary plasminogen activator, tPA, and PAI-1 (22). Importantly, the magnitude of differences with HT use was greater than that of differences in endogenous hormone levels.
We observed no significant association of endogenous E2 concentrations with factor VII-c concentrations, nor did we identify a difference in women who transitioned to postmenopause status compared with the premenopausal state. This later observation is in contradiction to the findings of Scarabin et al. (23) and other studies (24, 25, 26, 27) that show higher mean levels of factor VII-c and factor VII-a in postmenopausal women compared with premenopausal women. However, our longitudinal study compared the same women as they moved through the transitional states, which may lead to different conclusions than when comparing two different groups with differing menopausal states, who may have underlying intrinsic differences that account for the variation in factor VII-c values.
Baseline hemostatic factor profiles of subsequent HT users differed slightly, but importantly, from the profiles of endogenous hormone levels. Thus, HT use may prove inappropriate as a model for endogenous hormone concentrations among women in midlife. Although adoption of HT use was associated with suppressed PAI-1 concentrations, hsCRP concentrations were markedly higher among those adopting HT use. Studies of both oral contraceptives and HT products have reported selective elevations of CRP (28, 29), suggesting that estrogens may have a direct effect on the liver and increase CRP levels. Other studies have noted that this elevated hsCRP response to HT may be limited to a specific set of acute phase reactants and may not occur with other acute phase proteins, such as serum amyloid A (30).
Higher FSH concentrations were associated with higher factor VII-c and tPA levels and were negatively associated with hsCRP and fibrinogen concentrations. Although there is little research to guide the interpretation of these findings, there are a number of possible explanations. First, these findings may reflect the aging process. Alternatively, FSH and IGF-I have been reported to induce the accumulation of low-density lipoprotein receptor mRNA in granulosa cells (31); hence, FSH has been shown to have the potential for genomic activity that could be associated with an inflammatory response. Thus, the FSH relationships may reflect an underevaluated physiological response, but currently there is no ready explanation.
This study includes strengths and limitations. The study incorporates the direct measurement of endogenous hormones annually, assessment of menopause status annually, and adoption of exogenous hormone replacement use. This study involves a large sample being followed across time in the age range where the marked gender differences in atherosclerosis and coronary artery disease have been documented (3). Furthermore, members of the group were either premenopausal or early perimenopausal (based on menstrual bleeding definitions) and were not using HT at study onset, helping to establish the temporality of events. Despite these marked strengths, there are limitations. Obviously, the age and size of the population are insufficient to generate hard cardiovascular end points. Those outcomes will only become available after long-term follow-up of this cohort. It is recognized that the endogenous hormones described in this study are limited in their ability to represent the bioavailable fraction of the hormone that can potentially cross cell membranes and bind to nuclear steroid receptors (32), and that annual specimen collection, even when timed to a day of the cycle window and a time of day, must be interpreted with caution given the cyclic and pulsatile nature of circulating endogenous hormones.
In summary, the roles of endogenous hormones and the hemostatic risk factors for heart disease in women have not been adequately evaluated. Most studies have not assessed endogenous estrogens in relation to hemostatic factors, but focused instead on oral contraceptive use, HT use, or cross-sectional comparisons of pre- vs. postmenopausal women. We found that lower endogenous E2 levels were associated with higher levels of PAI-1 and tPA, consistent with a mechanism of greater clearance of fibrinolytic factors with higher endogenous E2 levels. There were no significant associations of E2 and fibrinogen, factor VII-c, or hsCRP. Menopause status, defined by regularity and frequency of menstrual bleeding in the time interval between annual examinations, was not associated with hemostatic risk factors. Finally, HT use was not a good proxy for the associations of endogenous hormone concentrations and hemostatic factors among premenopausal women or transitioning women. Depending upon the formulation, HT may reduce CVD risk from the fibrinolytic component but increase risk in the inflammatory component (hsCRP).
|
Footnotes |
|---|
First Published Online August 16, 2005
Abbreviations: BMI, Body mass index; CRP, C-reactive protein; CV, coefficient of variation; CVD, cardiovascular disease; E2, estradiol; hsCRP, human serum CRP; HT, hormone therapy; PAI-1, plasminogen activator inhibitor-1; SI, Systeme International; tPA, tissue plasminogen activator.
Received March 17, 2005.
Accepted August 9, 2005.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
