University of Southern California Dissertations and Theses

Associations between sex steroid hormones, hemostatic factors and atherosclerosis

(USC Thesis Other)

Associations between sex steroid hormones, hemostatic factors and atherosclerosis

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content ASSOCIATIONS BETWEEN SEX STEROID HORMONES, HEMOSTATIC
FACTORS AND ATHEROSCLEROSIS
by
Cheryl Vigen
A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(EPIDEMIOLOGY)
August 2007
Copyright 2007 Cheryl Vigen
ii
Table of Contents
List of Tables iii
List of Figures v
Abbreviations vi
Abstract viii
Introduction: 1
Chapter 1: Literature Review – Associations of Fibrinogen, D-dimer, 9
factor VII, tPA and PAI-1 with Atherosclerosis
Chapter 2: First Data Analysis – Postmenopausal Oral Estrogen Therapy 36
Affects Hemostatic Factors, but does not Account for Reduction
In the Progression of Subclinical Atherosclerosis
Chapter 3: Second Data Analysis - Levels of Circulating Hemostatic
Factors Are Related To Postmenopausal Estrogen Therapy, But
Not To Circulating Estrogen Levels 56
Chapter 4: Grant Proposal – Hemostatic Supplement to the ELITE Trial 76
Bibliography 138
iii
List of Tables
Table 1: Associations of Hemostatic Variables with Cardiovascular Disease 2
Table 2: Associations of Hormone Therapy with Hemostatic Variables 4
Table 3: Studies Relating Fibrinogen to Atherosclerosis 16
Table 4: Studies Relating D-dimer, factor VII, and PAI-1 to Atherosclerosis 23
Table 5: Baseline Characteristics of EPAT Participants with Hemostatic
Measures 42
Table 6: Baseline and Mean On-trial Values of Hemostatic Variables 44
Table 7: Rate of Change in CIMT by Hemostatic Factors - Estradiol and
Placebo Groups Combined 48
Table 8: Rate of Change in Hemostatic Factors by CIMT with Estradiol and
Placebo Groups Combined 49
Table 9: Baseline Characteristics of EPAT Participants with Hormone and
Hemostatic Measures 61
Table 10: Baseline and Mean On-trial Levels of Hormone and Hemostatic
Variables 62
Table 11: Association of Mean On-trial Hemostatic Factors with Mean On-trial
Hormone Levels Adjusted for age and BMI 65
Table 12: Association of Mean End of Trial Hemostatic Factors with Mean
On-trial Hormone Levels adjusted for age and BMI 67
Table 13: Association of Mean On-trial Hemostatic Factors with Mean On-trial
Androgens and SHBG Levels Adjusted for Estradiol Levels 69
Table 14: Known Effects of Postmenopausal HT 90
Table 15: Hypothesized Effects of Postmenopausal HT 91
Table 16: On-trial Comparison of Medication Compliance by Treatment Group 101
in EPAT
iv
Table 17: Comparison of CIMT by Treatment Group in EPAT 101
Table 18: On-trial Comparison of Medication Compliance by Treatment Group 102
in WELL-HART
Table 19: Baseline and Mean On-trial Values of Hemostatic Variables in EPAT 104
Table 20: Laboratory Variables Funded in ELITE 108
Table 21: Laboratory Measurements Proposed for this Supplement 108
Table 22: Detectible Treatment Group Differences within menopause strata 114
for Alpha=0.05 and 80% Power Assuming 10% Dropout Rate
Table 23: Correlation Coefficients between Hemostatic Factors and Sex 115
Steroid Hormones Observed in EPAT
v
List of Figures
Figure 1: Mechanisms for Associations of Hormone Therapy with Hemostatic 5
Factors
Figure 2: Change in tPA over Time by Treatment Group 45
Figure 3: Change in Factor VII over Time by Treatment Group 45
Figure 4: Change in D-dimer over Time by Treatment Group 46
Figure 5: Change in Albumin over Time by Treatment Group 46
Figure 6: Least Squares Means of On-trial CIMT Adjusted for Age and BMI 47
by Quartiles of Mean On-trial tPA (ng/mL)
Figure 7: The Coagulation and Fibrinolytic Systems 90
vi
Abbreviations
BMI – Body Mass Index
BVAIT - B-Vitamin Atherosclerosis Intervention Trial
CAC – Coronary artery calcium
CAD – Coronary artery disease
CHD - Coronary Heart Disease
CI – Confidence interval
CIMT – Carotid intima media thickness
CRP – C-reactive protein
CVD – Cardiovascular disease
DBP – Diastolic blood pressure
DHEA – Dehydroepiandrosterone
ELITE – Early versus Late Intervention Trial with Estrogen
EPAT – Estrogen in the Prevention of Atherosclerosis Trial
ET – Estrogen therapy
GFC – Global fibrinolytic Capacity
HR - Hazard Ratio
HT – Hormone therapy
MI – Myocardial infarction
MPA - medroxyprogesterone acetate
PAI-1 – Plasminogen activator inhibitor type 1
SBP – Systolic blood pressure
vii
SD – Standard deviation
SE – Standard error
SFM – Soluble fibrin monomer
SHBG – Sex hormone binding globulin
tPA – Tissue plasminogen activator
VEAPS - Vitamin E Atherosclerosis Prevention Study
vWF – von Willebrand factor
WELL-HART – Women's Estrogen-Progestin Lipid-Lowering Hormone
Atherosclerosis Regression Trial
WISH - Women’s Isoflavone Soy Health trial
viii
Abstract
Numerous cardiovascular risk factors have been identified, and lifestyle and drug
interventions are available for modification of many of these risk factors. Nevertheless,
cardiovascular disease remains the leading cause of death among both men and women
in the industrialized world. This has spurred interest in identifying novel risk factors
and markers in order to reduce cardiovascular disease morbidity and mortality.
Chapter 1 of this dissertation explores existing knowledge regarding hemostatic factors
(factor VII, fibrinogen, tPA, PAI-1 and D-dimer), and how they relate to atherosclerosis
measured by angiography, ultrasonography, computed tomography, or ankle/brachial
blood pressure ratio. Study results from 23 relevant, published, original research papers
were summarized and discussed.
Data analyses based on the Estrogen in the Prevention of Atherosclerosis Trial (EPAT)
are included. Chapter 2 analyzes the effect of postmenopausal hormone therapy on
levels of hemostatic factors (those listed above, plus albumin) and the relationship
between these factors and progression of subclinical atherosclerosis in 186
postmenopausal women. Significant treatment effects on hemostatic factors were
found, but no hemostatic factor was associated with CIMT progression.
Chapter 3 compares circulating levels of estrone, estradiol, free estradiol, testosterone,
free testosterone, androstenedione, DHEA and SHBG in 179 women from EPAT.
ix
Levels of tPA, PAI-1, albumin, and factor VII were related to circulating levels of
estrogens in a combined sample of treated and untreated women, but associations were
not independent of treatment group. tPA and PAI-1 were related to circulating SHBG
independent of treatment group.
Chapter 4 of this dissertation is a grant proposal, Hemostatic Supplement to the ELITE
Trial. The Early versus Late Intervention Trial with Estrogen (ELITE) is a randomized,
double-blind, placebo-controlled trial of 17- estradiol. This supplement will study the
effect of hormone therapy on hemostatic factors (factor VII, fibrinogen, tPA antigen,
tPA activity, PAI-1, soluble fibrin monomer, global fibrinolytic capacity, and von
Willebrand factor), the associations between circulating hormones and hemostatic
factors, and the relations between hemostatic factors and atherosclerosis. It will analyze
how these relations differ in four subgroups of women studied in ELITE, and will add
to our understanding of the role of hemostatic factors in atherosclerosis.
1
INTRODUCTION
Several modifiable cardiovascular risk factors have been identified (for example, high
LDL, low HDL, high triglycerides, hypertension, low physical activity, smoking,
diabetes and obesity) and both lifestyle and drug interventions are available for the
modification of these risk factors. Despite this, cardiovascular disease remains the
leading cause of death among both men and women in the industrialized world. This
has spurred an interest in the identification of novel risk factors and their markers in an
attempt to reduce morbidity and mortality associated with cardiovascular disease.
Hormone Therapy and Cardiovascular Disease
Postmenopausal hormone therapy (HT) and its relation to cardiovascular disease is an
area that has attracted much recent attention. Findings from the Women’s Health
Initiative seem to have resulted in a rapid decline in prescriptions for postmenopausal
HT, but not all treatment regimens were affected equally, and even after this decline,
approximately 57 million prescriptions were written for 10 million postmenopausal
women in 2003
42
. This indicates that women and their physicians want the benefits of
HT but are concerned about possible risks.
To determine which populations are favorably or adversely affected by estrogen
therapy, understanding the complex biological effects of estrogen is required. Estrogen
is associated with decreased LDL-cholesterol and increased HDL-cholesterol
29
; the
protective effect of estrogen on atherosclerosis is greater than explained by the effect of
2
estrogen on lipids
49
. Studies finding no clinical benefit with estrogen therapy suggest
improved lipid profiles do not necessarily reduce cardiovascular risk. Hemostatic
factors may be non-lipid mechanisms through which estrogen therapy affects
cardiovascular risk.
Hemostatic Factors and Cardiovascular Disease
Hemostatic factors, including both coagulation and fibrinolytic factors, have been
recently evaluated in relation to cardiovascular risk. In particular, fibrinogen as a risk
factor for cardiovascular events has been studied at least since the 1950s
61
. The
identification of a hemostatic factor as a risk factor for cardiovascular events, however,
does not detail the mechanism of the association. While it is possible that the
hemostatic variable may act directly on the initiation and progression of atherosclerosis,
its effects may also involve other factors predisposing to cardiovascular events, such as
the composition of the atherosclerotic plaque and its vulnerability to rupture.
Table 1. Associations of Hemostatic Variables with Cardiovascular Disease
Positive Negative Unknown
Fibrinogen Albumin GFC
Factor VII
tPA antigen
PAI-1
D-dimer
SFM
Von Willebrand Factor
__________________________________________________________________
While many studies have demonstrated a relationship between hemostatic factor levels
and cardiovascular disease (Table 1), it has not been clearly demonstrated that the
3
relationship acts through atherosclerosis. However, a rationale for this hypothesis is
that fibrin is found in atherosclerotic plaque
103
; excessive fibrin production may
contribute to elevated atherosclerosis build-up. Factor VII and fibrinogen are
components of the coagulation system that are associated with increased cardiovascular
risk
2,20,47,48,68,70,112
. However, high levels of these factors do not necessarily cause fibrin
production. High levels of factor VII cannot cause coagulation in the absence of tissue
factor, while high levels of fibrinogen will not result in fibrin formation in the absence
of thrombin
25
. In addition, low levels of circulating fibrinogen do not necessarily
indicate reduced coagulation potential since hepatic fibrinogen synthesis can increase
many times over in acute-phase response
13
. Nevertheless, increased levels may be
markers for an activated coagulation system. SFM is further down the coagulation
pathway than factor VII or fibrinogen and is a measure of the amount of fibrin
generated
74
.
The effectiveness of the fibrinolytic system in removing fibrin from a coronary artery
endothelial lesion may be equally or even more important than the amount of fibrin
produced in preventing cardiovascular events or atherosclerotic build-up. tPA converts
plasminogen to plasmin which dissolves the fibrin clot
13,50
while PAI-1 inhibits tPA. D-
dimer is a measure of the amount of fibrin degradation product
25
, and GFC is a measure
of the amount of D-dimer generated in a plasma sample when a standardized clot is
degraded
116
.
4
Sex Steroid Hormones and Hemostasis
Postmenopausal therapy with 17-estradiol increases levels of circulating total
estradiol, free estradiol, estrone and SHBG (see Chapter 4). Although total testosterone
levels are not changed by postmenopausal estradiol treatment, the increase in SHBG
causes a decrease in free testosterone. Therefore, the effects of postmenopausal
hormone therapy on hemostasis must be viewed in relation its effects on a hormonal
profile and not to circulating estradiol alone.
Table 2. Associations of Hormone Therapy with Hemostatic Variables
Positive Negative Unknown or Inconsistent
Factor VII Fibrinogen D-dimer
GFC tPA antigen SFM
tPA activity PAI-1
Albumin
__________________________________________________________________
Associations of postmenopausal hormone therapy with hemostatic factors are
summarized in Table 2. One mechanism for the associations may be related to hepatic
effects, especially with respect to variables produced in the liver (factor VII, fibrinogen,
PAI-1, albumin and SHBG). A second may be related to effects of the vascular
endothelium, which produces tPA, vWF and possibly PAI-1, and expresses both
estrogen and androgen receptors and thus may be sensitive to both levels of estrogen
and testosterone. These mechanisms are illustrated in Figure 1. In addition, it has been
shown that at the cell membrane level SHBG has an effect on transcriptional activity of
steroid hormone receptors, and is not only involved in the transport of hormones
88
.
5
Figure 1. Mechanisms for Associations of Hormone Therapy with Hemostatic Factors.
Dissertation Outline
This dissertation will analyze the associations among hemostatic factors, sex steroid
hormones, exogenous estrogen, and atherosclerosis.
Chapter 1
The first part of the dissertation will explore the existing knowledge regarding the
coagulation variables factor VII and fibrinogen and the fibrinolytic variables tPA, PAI-1
and D-dimer, and how these hemostatic factors relate to atherosclerosis measured by
HT
First pass liver Vascular endothelium
Produced
in the
Produced by the
endothelium:
Affect levels
of:
D-dimer
SFM
GFC
Factor VII
Fibrinogen
PAI-1
tPA
PAI-1
von Willebrand factor
SHBG
Albumin
6
angiography, ultrasonography, computed tomography, or ankle/brachial blood pressure
ratio. The literature review begins with a brief summary of the hemostatic variables and
the atherosclerosis imaging technology used in research. Twenty-three relevant
published papers of original research were identified using Ovid searches and review of
the reference lists in related papers. Study results were summarized and discussed.
This work is presented in Chapter 1 of this dissertation.
Chapters 2 and 3
The second part of the dissertation will present two data analysis papers. Both papers
use data from the Estrogen in the Prevention of Atherosclerosis Trial (EPAT), a
randomized, double-blind placebo-controlled clinical trial that tested if 17-estradiol
would reduce the progression of subclinical atherosclerosis in healthy postmenopausal
women. The first of these, titled Postmenopausal Oral Estrogen Therapy Affects
Hemostatic Factors, but does not Account for Reduction in the Progression of
Subclinincal Atherosclerosis, has been accepted for June, 2007 publication in the
Journal of Thrombosis and Haemostasis. This paper explores the relationship between
plasma levels of factor VII, fibrinogen, tPA, PAI-1, D-dimer and albumin, and
atherosclerosis measured with B-mode ultrasonography of the carotid artery.
Hemostatic factors, albumin and carotid atherosclerosis were all measured as part of
EPAT. EPAT data had previously been analyzed for the association between estrogen
treatment and progression of subclinical atherosclerosis. This paper contains the first
analysis of the hemostatic data collected as part of this trial.
7
The second data analysis paper, Levels of Circulating Hemostatic Factors Are Related
To Postmenopausal Estrogen Therapy, But Not To Circulating Estrogen Levels,
analyzes plasma levels of hemostatic variables factor VII, fibrinogen, tPA, PAI-1, and
D-dimer, and their associations with the sex steroid hormones estrone, estradiol, free
estradiol, testosterone, free testosterone, androstenedione and DHEA and the hormone-
binding protein SHBG. Data for this analysis also arise from the EPAT trial and the
hormone-related variables were all obtained as part of that trial. This paper is the first
analysis of the hemostatic factors in relation to the circulating hormone levels.
Chapter 4
The third part of the dissertation is a grant proposal, Hemostatic Supplement to the
ELITE Trial. The Early-Late Intervention Trial of Estrogen (ELITE) is a randomized,
double-blind, placebo controlled trial of 17- estradiol that is currently enrolling
patients. This trial will study the effect of estrogen on the progression of subclinical
atherosclerosis according to whether the woman is in recent or late menopause. As
shown in the first data analysis paper, postmenopausal estrogen use is associated with
changes in the levels circulating hemostatic factors. However, there is evidence that
vascular response to estrogen varies by time since menopause, and thus the effect of
estrogen on progression of subclinical atherosclerosis may also vary by time since
menopause. The effect of hemostatic factors on subclinical atherosclerosis is mixed,
which may be partly due to the fact that populations differing in their existing
atherosclerosis status may respond differently to levels of circulating hemostatic factors.
In addition, vascular response to estrogen may affect either the response of the
8
vasculature to hemostatic factors or may affect the production of hemostatic factors,
especially those produced in the vascular endothelium. Determining the effect of HT on
hemostatic factors, the associations between circulating sex-steroid hormones and
hemostatic factors, and the relations between hemostatic factors and atherosclerosis in
addition to how these relations may differ in the four subgroups of women studied in
ELITE (early/late postmenopausal and estrogen/placebo treatment) will add to our
understanding of the role of hemostatic factors in atherosclerosis.
B-mode ultrasonography of the carotid artery is already being performed every 6
months as part of the base ELITE trial. Plasma and serum samples are being collected
every 6 months and are being stored for the potential determination of levels of
circulating hormones and hemostatic variables factor VII and tPA antigen. This
proposed supplement will include the analysis of baseline and on-trial stored blood
samples for factor VII and tPA antigen, and end-of-trial analysis of blood samples for
fibrinogen, tPA activity, PAI-1, soluble fibrin monomer (SFM), global fibrinolytic
capacity (GFC) and von Willebrand factor (vWF). The use of much data that is already
being collected as part of the base trial will provide a very economical study of
hemostatic variables and their relationships to atherosclerosis and circulating hormones
according to duration since menopause.
9
CHAPTER 1
LITERATURE REVIEW
ASSOCIATIONS OF FIBRINOGEN, D-DIMER, FACTOR VII,
tPA, AND PAI-1 WITH ATHEROSCLEROSIS
Introduction
Several modifiable cardiovascular risk factors have been identified (for example, high
LDL, low HDL, high triglycerides, hypertension, low physical activity, smoking,
diabetes and obesity) and both lifestyle and drug interventions are available for the
reduction of these risk factors. Despite this, cardiovascular disease remains the leading
cause of death in the industrialized world. This has spurred an interest in the
identification of novel risk factors and their markers in an attempt to reduce morbidity
and mortality associated with cardiovascular disease.
Hemostatic factors, including both coagulation and fibrinolytic factors, have been
recently evaluated in relation to cardiovascular risk. In particular, fibrinogen as a risk
factor for cardiovascular events has been studied at least since the 1950s
61
. While it is
possible that the hemostatic variable may act on the initiation and progression of
atherosclerosis, its effects may also involve other factors predisposing to cardiovascular
10
events, such as the composition of the atherosclerotic plaque and its vulnerability to
rupture.
Arterial imaging technologies such as angiography, intravascular ultrasound, CT-
imaging, and ultrasonography of the carotid artery have been used in studies focused on
factors related to the presence, extent, and progression. While fibrinogen is an obvious
risk factor candidate because of its presence in atherosclerotic lesions
103
, the
associations between fibrinogen and other hemostatic factors have lead to investigations
of other hemostatic variables in relation to atherosclerosis.
This paper will review studies that have investigated the relationships between various
hemostatic factors and atherosclerosis measured by angiography, ultrasonography,
computed tomography, or ankle/brachial blood pressure ratio. The review will focus on
the specific hemostatic factors for which samples were collected in the Estrogen in the
Prevention of Atherosclerosis Trial (EPAT)
44
, namely, fibrinogen, tissue plasminogen
activator (tPA), plasminogen activator inhibitor-1 (PAI-1), factor VII, and D-dimer.
Hemostatic Factors
Fibrinogen (or Factor I) is a plasma protein produced in the liver that is the precursor
of fibrin in the coagulation process. It is higher in women than men, especially in
postmenopausal women, higher in smokers than nonsmokers, and increases with age
and body mass index
25,103
. Fibrinogen is reduced with postmenopausal estrogen
11
therapy
5,19
. Fibrinogen has been found in atherosclerotic lesions
103
and high plasma
levels have been linked with cardiovascular events in numerous studies
2,20,47,48,68,70,112
.
Factor VII (or stable factor or proconvertin) is part of the extrinsic coagulation system
which converts prothrombin to thrombin by activating factor X, which ultimately leads
to the conversion of fibrinogen to fibrin
25
. Factor VII levels increase with
postmenopausal estrogen therapy
19,73
. High values of factor VII have also been linked
to increased risk for cardiovascular events
70
.
tPA antigen is produced by vascular endothelial cells and activates fibrinolysis
13,50
. As
a fibrinolytic factor, tPA would be expected to be cardioprotective. However, most of
the tPA antigen measured in blood is in an inactive form bound with PAI-1
14
. In fact,
tPA antigen’s high correlation with PAI-1 activity levels, makes tPA antigen levels
essentially a surrogate measure of PAI-1
83
, i.e., high values of tPA antigen are
associated with low tPA activity. Consequently, plasma levels of tPA antigen have
been positively associated with cardiovascular events
36,47,69,81,84,90,102
, but none of these
studies adjusted for PAI-1. tPA activity has been measured in some atherosclerosis
studies
27,75,97
, and it is important to remember that its effect on atherosclerosis is
expected to be opposite that of tPA antigen. tPA antigen levels decrease with
postmenopausal estrogen therapy
56,81
PAI-1, which inhibits the action of tPA, is produced by the liver, adipose tissue, and
possibly endothelium
13,71
and is present in plasma, platelets, and extracellular fluid
103
.
12
Plasma levels of PAI-1 increase with age and BMI
103
and have been positively
associated with cardiovascular events
26,35,47,69,90,102
. PAI-1 levels decrease with
postmenopausal estrogen therapy
12,19
D-dimer is a fibrin degradation product
25
. Higher levels of D-dimer have been
associated with increased risk for cardiovascular events
26,56,81,90
. High values of this
factor could be indicative of high levels of fibrin production, making it a potential
atherogenic risk factor; this would consistent with its relationship to cardiovascular
disease risk. On the other hand, high values could also be indicative of efficient
fibrinolysis. Estrogen therapy does not appear to affect D-dimer levels
19,56,81
Atherosclerosis Imaging Technologies
Angiography is the gold standard for imaging coronary arteries but its use in research
is seriously limited by its invasiveness, such that angiographic studies can only be
performed among subjects with possible clinical manifestations of cardiovascular
disease
39
. Thus the population available for research with coronary angiography is
limited, and generalization of results in these studies is questionable. A number of
measures of atherosclerosis determined by angiography have been used, including
average and maximal percent stenosis
23,27,37
and minimum lumen diameter of coronary
artery lesions
9
, specialized angiographic scoring systems
7,63,93,97
, and expert panel
evaluations
37,94
.
13
Intravascular Ultrasonography is also an invasive procedure, but may provide some
information not available by angiography
117
. In particular, images from intravascular
ultrasound can differentiate the media from the intima and adventitia in muscular
arteries and can help define the plaque boundaries. The vessel wall can be viewed in all
directions, its thickness can be measured, and the nature of lesions can be better defined
as soft, fibrous, or calcified. Measures obtained using this imaging method are lumen
cross-sectional area, plaque volume and blood flow
75
.
B-Mode Ultrasonography is a non-invasive procedure used to obtain two-dimensional
images of shallow vessels. B-mode ultrasonography has been used to measure
atherosclerosis in the carotid
3,10,16,33,51,58,89-91,113
, femoral
57,58
and abdominal arteries
58
.
Ultrasonography of the carotid artery has been extensively used in clinical trials and
observational studies to measure atherosclerosis, and its association with coronary
atherosclerosis is well established
65
. It is particularly useful in measuring sub-clinical
atherosclerosis in studies where the use of an invasive imaging method is not ethically
possible (i.e., in persons without clinical indications of CVD). The ability to measure
sub-clinical atherosclerosis by ultrasonography has therefore allowed trials to be
performed and generalized in a much wider population. Importantly, this methodology
used in observational studies has allowed population-based and other samples of the
general population to be studied in relation to determinants of atherosclerosis. The
primary B-mode ultrasound measure of atherosclerosis is the intima-media thickness
(IMT), with larger values indicating greater atherosclerosis.
14
Electron Beam Computed Tomography (ECBT), while not as invasive as
angiography or intravascular ultrasound, involves a radiation dose equivalent to, or even
greater than, that for coronary angiography and 8 to15 times greater than that received
during a chest x-ray
110
. ECBT is also considerably more expensive and far less portable
than B-mode ultrasonography. Therefore its use in large cohort studies is limited. The
atherosclerosis measure obtained with this method is the amount of aortic and/or
coronary artery calcification. However, it is not clear that the amount of calcification
correlates well with the risk of incident coronary heart disease, especially in low-risk
populations
110
, or in those with diabetes
82
, since calcified plaques may be more stable
than non-calcified plaques.
Ankle/Brachial Blood Pressure Ratio is measured as the ratio of arterial systolic
blood pressure measured in the ankle compared to that measured in the arm. A lower
ratio is indicative of more peripheral atherosclerosis with ratios less than 0.7 or 0.9
being indicative of peripheral arterial disease and considered risk factors for coronary
heart disease
1
. This simple, non-invasive, inexpensive measurement tool has been
found to be inversely related to ultrasound measures of carotid artery atherosclerosis
118
.
15
Studies Relating Hemostatic Factors to Atherosclerosis
Fibrinogen
19 studies
30, 31, 39, 40, 42-45, 47-54, 56, 62, 63
were identified that related levels of fibrinogen to
atherosclerosis (Table 3). These studies have varied widely in size (n=38-1902),
population (gender, age, community-based vs. CVD patients), and method used to
measure atherosclerosis. All but two studies
10,93
have shown a significant positive
association between atherosclerosis and high fibrinogen levels in at least one portion of
their analyses. In no case was fibrinogen found to be negatively associated with
atherosclerosis.
Five studies included more than 500 subjects. The largest study
51
measured fibrinogen
levels in patients comparing those with at least 50% stenosis determined by
ultrasonography of the carotid artery (n=318) to community-based volunteers who were
free of ischaemic cerebrovascular or heart disease (n=1584). The community-based
comparators were not ultrasonographically examined. In this study, fibrinogen levels
were significantly higher in the patients with stenosis (4.7 vs. 3.1 g/L, p<.0001) after
adjustment for age, sex, BMI and smoking. However, these findings may suffer from
subject selection bias, as elevated fibrinogen levels may have been associated with
atherosclerosis as well another factor associated with the reason for the ultrasound
referral.
Table 3. Studies relating fibrinogen to atherosclerosis.
Study
Author
Year
Pub
Artery
Imaged
Atherosclerosis
Definition
Population N Results Adjustments
Chapman
16
2004 Carotid IMT determined by
ultrasonography
Australian community
based, balanced by gender,
ages 27-77
1111 Significantly higher in subjects
with greater IMT (p<.0001), but
not significant when adjusted
for established cardiovascular
risk factors.
Age, sex, waist-hip
ratio, LDL,
smoking,
hypertension, MI,
diabetes
>50% stenosis
determined by
ultrasonography
Danish patients with >50%
stenosis determined by
carotid ultrasonography
community based controls,
65% male
1902 Significantly higher in cases
(p<.0001).
BMI, smoking. Age
and sex matched
Kofoed
51
2003 Carotid
Echolucent (rupture-
prone) vs. echo-rich
(stable) plaques
determined by
ultrasonography
Danish patients with >50%
stenosis determined by
carotid ultrasonography,
65% male
318 Not significant. Age, sex, BMI,
smoking
Kullo
55
2003 Coronary Coronary artery calcium
determined by electron
beam computed
tomography
Non-Hispanic white,
hypertensive, 42% male,
mean age 66
354 Significantly higher in women
with higher CAC (p<.0001).
Not related in men.
1) None, 2)
“conventional risk
factors”, BMI, statin
and estrogen use
Auer
7
2002 Coronary 3-vessel, 27-point
angiographic score
categorized as 0-3, 4-8
or 9-27
White, consecutive
angiography patients, 59%
male, ages 38-84
100 Significantly higher in subjects
with moderate or severe
atherosclerosis (p<.01).
None
Bielak
8
2000 Coronary Calcium determined by
electron beam computed
tomography >80
th
percentile
Epidemiology of Coronary
Artery Calcification study –
community based, 50%
male, ages 40-69
227 Significantly higher in cases
(p<.01). Significantly higher in
female but not male cases after
adjustment for established
cardiovascular risk factors
(p<.0001).
Age, BMI, total
cholesterol, HDL,
smoking, SBP, CRP
17
Table 3 (continued). Studies relating fibrinogen to atherosclerosis.
Study
Author
Year
Pub
Artery
Imaged
Atherosclerosis
Definition
Population N Results Adjustments
>1 stenosis >25%
determined by
angiography
Significantly higher in cases
(p=.008).
None Eichner
23
1996 Coronary
Number of vessels
with >25% stenosis
White, female,
consecutive
angiography patients,
ages 48-61
101
Significantly higher in those with
greater disease.
Age
Levenson
58
1995 Carotid
Femoral
Abdominal
aorta
Number of plaques
identified by
ultrasonography
Men with
cardiovascular risk
factors, ages 40-60
652
Significantly higher in those with
more plaques (p<.01).
Multivariate regression
including variables
with p<.10: age,
triglycerides
Maximal CIMT
determined by
ultrasonography
Significantly positively correlated
in group 2 (p<.05). Non-
significantly positively correlated
in groups 1 and 3.
Agewall
3
1994 Carotid
3-level Plaque status
determined by
ultrasonography
Men, ages 57-77,
1) high risk with CVD,
2) high risk without
CVD, 3) low risk
180
Significantly higher in those in
groups 1 and 2 with higher plaque
levels (p=.012). Non-significantly
higher in those in group 3 with
higher plaque levels (p=.08)
None
Femoral
and/or
coronary
Ultrasonography
and/or angiography 36
Not presented. Lassila
57
1993
Ankle/brachial
pressure ratio
Stable peripheral
arterial occlusive
disease patients, 62%
male, ages 37-85
40
Inversely correlated with ratio
(p<.0002).
None
Willeit
113
1993 Carotid >66
th
percentile of
score of lesions
determined by
ultrasonography.
Community based, 51%
male, ages 40-79 909
Higher in elderly cases. Not
significant in middle-age cases
after adjustment, but for men it
was univariately significant.
Age and sex matched.
Stepwise regression:
Men: SBP, smoking,
diabetes; Women:
SBP, smoking,
apolipoprotein A-I
18
Table 3 (continued). Studies relating fibrinogen to atherosclerosis.
Study
Author
Year
Pub
Artery
Imaged
Atherosclerosis
Definition
Population N Results Adjustments
Bonithon-
Kopp
10
1991 Carotid 3 levels according to
IMT and plaque
determined by
ultrasonography
Healthy French women,
ages 45-54 517
Non significant increase with
increasing atherosclerosis (p=.27)
Age
Grotta
33
1989 Carotid Progression of
atherosclerosis
determined by
ultrasonography
Patients with
asymptomatic carotid
stenosis, 50% male,
ages 50-89
38
Baseline and follow-up values
significantly higher in progressors
Stepwise regression:
CAD, LDL
Handa
37
1989 Coronary Number of vessels with
>75% stenosis and
Gensini’s severity
score
Japanese angiography
patients, 71% male,
ages 25-82
229
Significantly higher in men (but
not women) with more stenoses
and in those with higher Gensini’s
scores (p<.01 for age only, or for
full adjustment).
1) Age, 2) Age,
hypertension, total
cholesterol, HDL,
smoking, alcohol, BMI
Abnormal angiograms 354 Significantly higher in those with
abnormal angiograms
Age Schneidau
94
1989 Carotid
Increased disease
measured by repeat
angiography
Patients investigated for
cerebral vascular
disease or ischemic
heart disease 209 Not significant Age
Francis
27
1988 Coronary 60% stenosis of at least
1 major artery
determined by
angiography
Angiography patients
127
Significantly higher in cases
(p<.05).
None
Salonen
91
1988 Carotid 4 levels according to
IMT, plaque and
stenosis determined by
ultrasonography
Eastern Finnish men,
ages 42-60
412 Higher in those with more
atherosclerosis (Adjustment 1)
p<.001, 2) p<.05, 3) p=ns)
1) None, 2) Age, 3)
Age, BMI, LDL,
HDL
2
, HDL
3
,
smoking, hypertension
19
Table 3 (continued). Studies relating fibrinogen to atherosclerosis.
Study
Author
Year
Pub
Artery
Imaged
Atherosclerosis Definition Population N Results Adjustments
Schmitz-
Huebner
93
1988 Coronary 5-level scale determined by
number of vessels with
stenoses<50% and >50%
Angiography patients,
80% male, mean age
52
225
Not significant Stepwise regression:
MI history, sex,
worsening angina,
triglycerides, ejection
fraction.
Number of major vessels
with >50% stenosis
Non-significantly higher in those
with greater scores.
Small
97
1987 Coronary
0-4 point score system
White, male,
angiography patients,
mean age 49
100
Significantly higher in those with
higher scores.
None
Lowe
63
1980 Coronary Stenosis of 2-3 major vessels
determined by angiography
Male angiography
patients, ages 30-55 50
Higher mean blood viscosity
partly due to higher fibrinogen
concentration in cases.
None
20
The second largest study
16
completed ultrasound examinations of the carotid artery on
1111 community-based volunteers. This study found that those with plaque, defined as
an increased IMT of at least 1mm, compared to those with no plaques, were more likely
to have high levels of fibrinogen (odds ratio=1.43 per g/L of fibrinogen, 95% CI =
[1.10,1.84]). However, after adjustment for age, sex, waist-hip ratio, LDL, smoking,
hypertension, history of myocardial infarction and diabetes the association between
atherosclerosis and fibrinogen was no longer evident.
The next two large studies used stepwise logistic regression to determine what variables
were significantly positively associated with ultrasonographically-measured
atherosclerosis. In one of these studies
113
, among subjects at least 65 years old, but not
for younger subjects, fibrinogen was a significantly associated with carotid
atherosclerosis. The stepwise model also included systolic blood pressure and smoking
for both men and women, diabetes for men, and apolipoprotein B for women. The other
study
58
included a group of 652 men with cardiovascular risk factors. Fibrinogen, as
well as age and triglycerides were positively associated with atherosclerosis measured
by presence/absence and by number of sites of plaque measured in the carotid and
femoral arteries and the abdominal aorta.
The final large study
10
included 517 healthy women with atherosclerosis categorized
into one of three levels according to carotid IMT and plaque determined by
ultrasonography. This study found an age-adjusted, non-significant increase in
fibrinogen levels with increasing atherosclerosis level.
21
Whether or not fibrinogen is an independent risk factor for atherosclerosis remains to be
determined. Six of the studies reviewed did not adjust for any confounders
3,7,27,57,63,97
,
three adjusted for age only
10,23,94
, six studies adjusted for several established
cardiovascular risk factors
8,16,37,51,55,91
(including combinations of age, sex, BMI, waist-
hip ratio, smoking, alcohol, blood pressure, lipids, diabetes, history of myocardial
infarction, C-reactive protein, statin and estrogen use), and four used stepwise
regression models
33,58,93,113
. While some studies showed significant adjusted
relationships between atherosclerosis and fibrinogen, others did not.
All but two of the studies were cross-sectional. In a very small progression study
33
subjects were followed for up to 30 months with periodic B-mode ultrasonography of
the carotid artery. Stenosis was defined as the sum of the thickest plaque on each side
of the artery wall divided by the wall-to-wall diameter of the artery. Mean percent
stenosis was defined as the mean of stenoses measured in the distal common, bulb and
proximal internal common carotid arteries. Progressing atherosclerosis was defined as
an increase in mean percent stenosis of at least 19% or an increase in a single region of
at least 23%. This study found that baseline and follow-up levels of fibrinogen were
higher in subjects with progressing (n=8) versus those with non-progressing (n=30)
atherosclerosis. The other progression study
94
(n=209) compared baseline and 2-year
carotid angiograms, and defined progression of atherosclerosis as a narrowing of a
previously normal carotid bifurcation, an increase in percent stenosis of at least 20%, or
the occlusion of a previously open vessel. This study found that baseline fibrinogen
levels were higher in patients with abnormal carotid angiograms than in those with
22
normal angiograms, but that increased carotid atherosclerosis measured by repeat
carotid angiography was unrelated to fibrinogen levels.
There is no subset of the population for whom an independent relationship between
atherosclerosis and fibrinogen has been clearly established, nor has a clear list of
confounders for the relationship between fibrinogen and atherosclerosis been identified.
23
Table 4. Studies relating D-dimer, factor VII, tPA, and PAI-1 to atherosclerosis.
Study
Author
Year
Pub
Artery
Imaged
Atherosclerosis Definition Population N Results Adjustments
D-dimer
Salomaa
90
1995 Carotid Cases with IMT>90
th
percentile; controls with
IMT<75
th
percentile
determined by
ultrasonography
ARIC sub-study,
community based, 61%
male, mean age 56
914 Unadjusted:
significantly higher in
cases (p<.05).
Adjusted: unrelated
1) None, 2) Age, SBP,
total cholesterol, aspirin
use, time of day, BMI,
smoking
Femoral
and/or
Coronary
Duplex ultrasonography
and/or angiography
36 Significantly higher in
those with more
severe disease (p<.01).
Lassila
57
1993
Ankle/brachial pressure ratio
Stable peripheral
arterial occlusive
disease patients, 62%
male, ages 37-85 40 Inversely correlated
with ratio (p<.0001).
None
Factor VII
Abnormal angiograms 354 Not significant Age Schneidau
94
1989 Carotid
Increased carotid disease
measured by repeat
angiography
Patients investigated
for cerebral vascular
disease or ischemic
heart disease
209 Not significant Age
Schmitz-
Huebner
93
1988 Coronary 5-level scale determined by
number of vessels with
stenoses<50% and >50%
Angiography patients
185
Not significant Stepwise regression: MI
history, sex, worsening
angina, triglycerides,
ejection fraction.
tPA Antigen
Salomaa
90
1995 Carotid Cases with IMT>90
th
percentile; controls with
IMT<75
th
percentile
determined by
ultrasonography
ARIC sub-study,
community based, 61%
male, mean age 56
914 Unadjusted:
significantly higher in
cases (p<.05).
Adjusted: unrelated
1) None, 2) Age, SBP,
total cholesterol, aspirin
use, time of day, BMI,
smoking
Paramo
76
2001 Coronary Stenosis >75% determined by
angiography
Stable angina patients,
80% male, mean age 59
60 Significantly higher in
cases (p<.03).
None
24
Table 4 (Continued). Studies relating D-dimer, factor VII, tPA, and PAI-1 to atherosclerosis.
Study
Author
Year
Pub
Artery
Imaged
Atherosclerosis
Definition
Population N Results Adjustments
tPA Activity
Number of major vessels
with >50% stenosis
Small
97
1987 Coronary
0-4 point score system
White, male,
angiography
patients mean age
49
100 Not significant None
Francis
27
1988 Coronary CAD determined by
angiography
Angiography
patients
127 Significantly lower in cases
(p=.002).
None
Newby
75
2001 Coronay Plaque burden in left
anterior descending
artery, determined by
intravascular ultrasound
Angiography
patients, 68%
male, mean age 56
25 Significantly lower in patients
with higher plaque burden
(p=.003).
None
PAI-1 Antigen
Salomaa
90
1995 Carotid Cases with IMT>90
th
percentile; controls with
IMT<75
th
percentile
determined by
ultrasonography
ARIC sub-study,
community based,
61% male, mean
age 56
910 Unadjusted: significantly
higher in cases (p<.05).
Adjusted: unrelated
1) None, 2) Age, SBP, total
cholesterol, aspirin use, time of
day, BMI, smoking
Sakata
89
2004 Carotid IMT quartile determined
by ultrasonography
Japanese
community based,
47% male, ages
34-91
522 Men: Adjustment 1) Not
significant, 2 & 3) Higher in
those with greater IMT
(p<.001); Women: Higher in
those with greater IMT
(Adjustment 1) p=.07, 2) p=.23,
3) p=.32)
1) None, 2) Age, 3) Age,
alcohol, smoking, BMI,
diabetes, hypercholesterolemia,
SBP, hypertensive drug use
25
Table 4 (Continued). Studies relating D-dimer, factor VII, tPA, and PAI-1 to atherosclerosis.
Study
Author
Year
Pub
Artery
Imaged
Atherosclerosis
Definition
Population N Results Adjustments
PAI-1 Activity
Francis
27
1988 Coronary CAD determined by
angiography
Angiography patients 127 Significantly higher
in cases (p<.001).
None
Maximal IMT
determined by
ultrasonography
Not significant Agewall
3
1994 Carotid
3-level Plaque status
determined by
ultrasonography
Men, ages 57-77,
1) high risk with
CVD,
2) high risk without
CVD,
3) low risk
180
Not significant
None
Paramo
76
2001 Coronary Coronary stenosis
>75% determined by
angiography
Stable angina
patients, 82% male,
mean age 59
60 Not significant, but
33% higher in cases.
None
26
D-dimer
Only two studies, one large (n=914)
90
, and one small (n=40)
57
were identified analyzing
the relationship between D-dimer and atherosclerosis (Table 4). The larger of these two
studies (n=914)
90
was a case-control study nested in the Atherosclerosis Research in
Communities (ARIC) cohort. Ultrasonography of the carotid artery was used to
determine cases (subjects with CIMT greater than the 90
th
percentile for the full cohort)
and controls (subjects with CIMT less than the 75
th
percentile). The smaller study
57
was
cross-sectional and included patients with stable peripheral arterial occlusive disease.
In this study, levels of D-dimer were related to ankle-brachial pressure ratio (n=40) and
to the extent of atherosclerosis determined by ultrasonography and/or angiography of
lower-limb arteries (n=36). In both the large and small studies, higher levels of D-
dimer were found to be a risk factor for atherosclerosis on univariate analysis. Only
ARIC provided an additional adjusted analysis and found that D-dimer and
atherosclerosis case-control status were not related when adjusted for age, systolic
blood pressure, total cholesterol, aspirin use, time of day, BMI, and smoking.
Factor VII
Only two studies
93,94
were found evaluating the relation between factor VII and
atherosclerosis (Table 2). One of these was a cross-sectional study of patients
undergoing coronary angiography (n=185)
93
and categorized the extent of coronary
atherosclerosis on a 5-level scale. The other studied patients evaluated for cerebral
vascular disease or ischemic heart disease, and presented two different analyses
94
. A
baseline cross-sectional analysis (n=354) used a 5-level scale of atherosclerosis of the
27
carotid artery determined by angiography. A 2-year progression analysis with repeated
carotid angiography (n=209)
94
defined carotid atherosclerosis progression as an increase
in percent stenosis of at least 20% or the occlusion of a previously open vessel. Factor
VII was not significantly related to atherosclerosis in either of these studies, either in
cross-sectional or longitudinal analyses.
tPA
Three studies measured tPA activity
27,75,97
and two measured tPA antigen
76,90
. All three
tPA activity studies included patients undergoing coronary angiography. One study
(n=100) determined level of atherosclerosis by the number of major coronary vessels
with at least 50% stenosis
97
. Another (n=127) defined cases as patients with one or
more coronary stenoses of at least 60%, and controls as patients with no stenoses
27
. The
last study (n=25) defined atherosclerosis by plaque burden (plaque volume divided by
length) in the proximal left anterior descending coronary artery, determined by
intravascular ultrasound
75
. tPA activity was significantly inversely related to coronary
atherosclerosis in two of the three studies
27,75
. However, neither of these studies
controlled for potential confounders. tPA activity was not related to atherosclerosis in
the third study
97
.
tPA antigen was analyzed in the case-control study nested in ARIC
90
(n=914). tPA
antigen was also analyzed in a cross-sectional study of patients with stable angina
(n=60) in whom extent of atherosclerosis was determined by coronary angiography.
tPA antigen was a univariately significant risk factor for atherosclerosis in both studies,
28
but the relationship with carotid atherosclerosis was not significant in the additional
analysis performed in ARIC
90
adjusting for age, systolic blood pressure, total
cholesterol, aspirin use, time of day, BMI, and smoking.
PAI-1
Five studies investigated the relationship between PAI-1 (three measuring activity
3,27,76
and two measuring antigen
89,90
) and atherosclerosis. One of the PAI-1 activity studies
included patients undergoing coronary angiography (n=127), and defined cases as
patients with one or more coronary stenosis of at least 60%, and controls as patients
with no coronary stenoses
27
. Another study included men (n=180) with no
cardiovascular risk factors, with cardiovascular risk factors but no manifest disease, and
with manifest cardiovascular disease. PAI-1 activity levels were analyzed in relation to
both maximal CIMT and to a 3-level plaque measure determined by carotid artery
ultrasonography. PAI-1 activity was also analyzed in a cross-sectional study of patients
with stable (n=60) in whom extent of atherosclerosis was determined by coronary
angiography. None of the studies of PAI-1 activity were adjusted for potential
confounders.
One of the studies of PAI-1 antigen was ARIC (n=910). The other was a Japanese
community-based study (n=522) which related quartiles of CIMT determined by
ultrasonography to levels of PAI-1 antigen. Both of these studies presented unadjusted
analyses and analyses adjusting for several cardiovascular risk factors.
29
The study of angiography patients
27
found PAI-1 activity levels to be significantly
higher (p<.001) in patients with coronary artery disease. In patients with stable
angina,
76
PAI-1 activity levels were 33% higher in patients with greater coronary
stenosis, but the association was not statistically significant. In the remaining study
3
,
PAI-1 activity was not significantly related to carotid atherosclerosis. In ARIC, PAI-1
antigen was significantly positively associated with carotid atherosclerosis in an
unadjusted analysis. A statistically significant univariate relationship between PAI-1
antigen and carotid atherosclerosis was found in women in the Japanese study. In
ARIC, the association between PAI-1 antigen and atherosclerosis was not significant
when adjusted for age, systolic blood pressure, total cholesterol, aspirin use, time of
day, BMI, and smoking. When the Japanese study adjusted for age, alcohol use,
smoking, BMI, diabetes, hypercholesterolemia, systolic blood pressure, and
hypertensive drug use, a significant relationship between PAI-1 antigen and
atherosclerosis was found in men but not women.
While none of the reviewed studies suggest that PAI-1 is protective against
atherosclerosis, it is not clear whether or not PAI-1 is positively related to
atherosclerosis. The two large community-based studies of PAI-1 antigen differ in their
results before and after adjustment for potential confounders. There is also evidence
that the effect of PAI-1 may vary by sex. Effect modification by ethnicity is also a
possible explanation of disparate results. As with other hemostatic factors, careful
attention needs to be paid to potential confounders and effect modifiers in order to
evaluate the association between PAI-1 and atherosclerosis.
30
Discussion
Although many studies have analyzed the relationships between hemostatic factors
(especially fibrinogen) and atherosclerosis, study limitations and inconsistent results
prevent us from being able to make unequivocal statements regarding the effect of any
of the reviewed hemostatic factors on atherosclerosis. Factors that could have
influenced these studies’ results include study design (longitudinal vs. cross-sectional),
the atherosclerosis measurement method, the population studied (gender, stage of
disease, population- versus clinic-based sampling), and whether or not analyses were
adjusted for other cardiovascular risk factors.
Only two studies
33,94
measured progression of atherosclerosis on longitudinal analysis.
One study measured atherosclerosis at baseline and at 2 years
94
, while the other
measured atherosclerosis at 2- to 6- month intervals over a 30-month study period
33
.
Although many of the other studies characterized patients as either cases or controls, all
were actually cross-sectional studies and the case and control terminology was used
only to identify patients who were above versus below a given dichotomous cutpoint of
the otherwise continuous measure of atherosclerosis. Whereas standard case-control
studies define cases as persons with an incident clinical diagnosis, the extension of this
design to a prevalent continuous outcome is questionable, and the interpretation of
measures of association are not as clear as for standard case-control design. In
particular, the temporal relationship between the exposure and the outcome is unclear.
31
Often there is strong biological rationale for causal direction when case-control study
designs are used. An association between hemostatic factors and atherosclerosis,
however, could be caused by atherosclerosis altering production (or clearance) of
hemostatic factors rather than hemostatic factors affecting the progression of
atherosclerosis. It is not unreasonable to hypothesize that atherosclerosis would affect
the levels of hemostatic factors produced in the vascular endothelium or contribute to
feedback that would affect hemostatic factor production in the liver.
In some studies more than two categories for the measure of atherosclerosis were used,
but the cutpoints were arbitrary. While the advantage of this is the clear separation of
atherosclerosis levels, it likely resulted in a loss of power over what could have been
obtained if atherosclerosis had been modeled as a continuous variable. In addition, the
arbitrary levels chosen for comparison have no clinical relevance.
Four studies
33,58,93,113
used a stepwise regression method to determine what variables
were associated with atherosclerosis. In such models, a variable may not meet the p-
value requirement for selection into the model, but nonetheless be a confounder. Such
stepwise approaches therefore may not include important confounders. Secondly, if
candidate variables for the model are highly correlated, one may be selected as
significantly related while the second variable is omitted even when there is no
difference between the relationships of each of the candidate variables with the outcome
variable. Stepwise regression models also tend to give misleading results when the
sample size is small or the number of candidate variables is large.
32
The two most common methods used to measure atherosclerosis in these studies were
coronary angiography and ultrasonography of the carotid artery. Angiography can only
be used in high risk patients leading to likely subject selection biases, while community-
based samples can be studied with carotid ultrasound. Consequently, the
atherosclerosis measured in the angiographic studies is likely to be more advanced than
that measured in ultrasonographic studies. Using angiography discounts potential
effects of hemostatic factors on early atherosclerosis. In addition, angiography is most
often indicated for patients with suspected high levels of atherosclerosis; patients with
lower suspected levels of atherosclerosis must have other indications for angiography
which may include risk factors associated with hemostatic variables.
The populations represented in these studies are highly variable, with different age
ranges and gender proportions, and most importantly, differing levels of existing
cardiovascular disease. Whether the hemostatic factors studied may be more or less
important in various stages of disease is unknown. Similarly, it is unknown whether
hemostatic factors may affect atherosclerosis differently by gender, race, age, smoking
status and other population characteristics.
Pearson has discussed the particular selection and measurement concerns that may
affect angiographic “case-control” studies
78
. Of most concern, because all subjects
must have some clinical indication for angiography, controls frequently have only
slightly less atherosclerosis than cases. Because an invasive procedure is required,
33
subjects are more likely to be patients exhibiting clinical cardiovascular symptoms,
excluding those with clinically silent but nonetheless high levels of atherosclerosis.
Studies are prevalence- rather than incidence-based, resulting in an underestimation of
risk factors important in the early stages of disease. In addition, symptomatic disease
may cause lifestyle or other changes that alter the measured levels of risk factors
responsible for pre-symptomatic disease. Biases may be introduced depending on the
criteria used to determine that a patient is at high enough risk to justify an invasive
diagnostic procedure. In addition, a physician’s decision to recommend a coronary
angiogram may be influenced by the patient’s risk factor profile, leading to numerous
selection biases that cannot be measured. The patient may choose whether or not to
seek medical treatment based on perceived risk. These biases in angiographic case-
control studies make it impossible to determine whether a variable is a risk factor for
atherosclerosis or an unrelated factor associated with symptoms precipitating
angiography. These biases are dramatically reduced when atherosclerosis is measured
by a non-invasive procedure that can be ethically performed on a general population
basis.
Several of these studies did not adjust for potential confounders. Generally these were
studies of patients undergoing coronary angiography, and thus a high level of
cardiovascular risk for the entire population could be assumed. However, adjustment
should include at the minimum age, gender, BMI and smoking status, variables known
to be associated with hemostatic variables and atherosclerosis. On the other hand, two
of the studies
16,93
went beyond reasonable adjustment and included history of MI as an
34
adjusting variable. Since MI is a manifestation and not a cause of atherosclerosis,
adjusting for it in a model attempting to identify atherosclerosis risk factors would tend
to attenuate the effect of any variable causing the atherosclerosis leading to the MI. For
example, the study by Schmitz-Huebner, et al.
93
is one of only two studies out of 19 that
found no relationship between fibrinogen and atherosclerosis. (It also found no
relationship between factor VII and atherosclerosis.) However, no univariate analysis
was presented; instead, a stepwise regression model was presented which included MI
history, gender, worsening angina, triglycerides and ejection fraction as significant
predictors of atherosclerosis. The insignificance of such established risk factors as age,
smoking status and cholesterol in this regression model further suggests that the
inclusion of MI history, worsening angina and ejection fraction (all manifestations
rather than causes of atherosclerosis) may be obscuring important atherosclerosis risk
factors.
Adjusting for age, gender and smoking status is appropriate since these variables are all
related to both atherosclerosis and to hemostatic factors. The relationships between the
hemostatic factors and many other atherosclerosis risk factors, such as cholesterol and
hypertension, are unclear. Thus it is unclear whether or not these are confounders that
should be included in models. Further studies evaluating the relationships between
hemostatic variables and recognized atherosclerotic risk factors need to be completed.
35
Conclusions
Unfortunately, few of the studies reviewed met the standards of adjusting at least for
age, sex and BMI, and not adjusting for cardiovascular events
8,37,51,55,89-91
. Among the
studies with appropriate covariate adjustment, all five studies of fibrinogen reported
significantly higher fibrinogen levels in subjects with greater atherosclerosis
8,37,51,55,91
.
Sometimes this relationship was found in men but not women
37
or the study only
included men
91
, while other studies reported the association in women but not men
8,55
.
These studies used a variety of imaging techniques and enrolled both healthy and
diseased populations. Further study of the relationship between fibrinogen and
atherosclerosis is clearly needed.
Only one D-dimer study
90
, one tPA study
90
, and two PAI-1 studies
89,90
met adjustment
standards. These four analyses were reported in two papers and studied population-
based subjects using carotid ultrasound measures of atherosclerosis. Only one
significant relationship was found: higher PAI-1 antigen in men (but not women) with
greater atherosclerosis. Neither factor VII study was appropriately adjusted. Further
study is necessary to confirm these results and to explore the relations, and possible
interactions, between hemostatic factors and other cardiovascular risk factors. Ideally,
these future studies would be longitudinal to help determine cause and effect
relationships, would be population-based (using a noninvasive imaging method) for
greater generalizability, and would explore a variety of covariates as potential
confounders and effect modifiers.
36
CHAPTER 2
FIRST DATA ANALYSIS
POSTMENOPAUSAL ORAL ESTROGEN THERAPY AFFECTS
HEMOSTATIC FACTORS, BUT DOES NOT ACCOUNT FOR REDUCTION IN
THE PROGRESSION OF SUBCLINICAL ATHEROSCLEROSIS
Introduction
While observational studies
32
suggest a protective cardiovascular effect of
postmenopausal estrogen therapy, in randomized clinical trials
43
the relationship is
more complex than previously thought. To determine which populations are favorably
or adversely affected by estrogen therapy, understanding the complex biological effects
of estrogen is required. Estrogen is associated with decreased LDL-cholesterol and
increased HDL-cholesterol
29
; the protective effect of estrogen on atherosclerosis is
greater than explained by the effect of estrogen on lipids
49
. Studies finding no clinical
benefit with estrogen therapy suggest improved lipid profiles do not necessarily reduce
cardiovascular risk. Hemostatic factors may be non-lipid mechanisms through which
estrogen therapy affects cardiovascular risk. From previous literature, we chose to
study hemostatic factors likely influenced by estrogen therapy and/or related to
atherosclerosis progression. The relationship of some hemostatic factors with
cardiovascular disease may act through acute thrombic events rather than long-term of
37
atherosclerotic plaque accumulation. Our chosen hemostatic factors had best evidence
as possible atherosclerosis risk factors.
Fibrinogen, a liver-derived protein, is the precursor of fibrin in the coagulation process.
It is higher in women than men, especially postmenopausal women
25,103
, but reduced
with postmenopausal estrogen therapy
19
. Fibrinogen is found in atherosclerotic lesions
103
; high levels are linked with cardiovascular events
20,51,70
.
Factor VII contributes to conversion of fibrinogen to fibrin
25
. High factor VII values
are linked to increased cardiovascular event risk
70
. Factor VII levels increase with
postmenopausal estrogen therapy
19
.
tPA, produced by vascular endothelial cells, activates fibrinolysis
13
. Most tPA antigen
in blood is in an inactive form bound with PAI-1
14
and is thus essentially a surrogate
measure of PAI-1
83
(high tPA antigen is associated with low tPA activity). tPA antigen
is positively associated with cardiovascular events
81,90
. tPA antigen levels decrease
with postmenopausal estrogen therapy
56,81
.
PAI-1 inhibits the action of tPA, is produced by the liver, adipose tissue, and possibly
endothelium
13
, and is present in plasma, platelets, and extracellular fluid
103
. PAI-1 is
positively associated with cardiovascular events
90
. PAI-1 levels decrease with
postmenopausal estrogen therapy
19
.
38
D-dimer, a fibrin degradation product
25
, is positively associated with cardiovascular
events
56,81,90
. Estrogen therapy does not appear to affect D-dimer levels
19,56,81
.
Albumin, the most abundant protein in blood, is a liver-derived binding and transport
protein
46
. Albumin binds numerous compounds (including sex steroid hormones and
long-chain fatty acids) reversibly and with great affinity
54
, making albumin a potential
factor in many processes including hemostasis. Albumin is negatively associated with
cardiovascular disease risk
20
.
Factor VII and fibrinogen may contribute to atherosclerosis evidenced by thrombic
deposits found in atherosclerotic lesions
38
. tPA and PAI-1 determine the fibrinolytic
potential of the circulatory system; balance toward fibrinolysis may prevent thrombi
incorporation into plaque and conversion to fibrous tissue. High D-dimer values could
result from fibrin degradation following excessive fibrin production, making it a
potential atherogenic risk marker. The negative association of albumin with
cardiovascular disease risk may be related to its binding and transport of molecules
involved in the atherosclerotic process.
The Estrogen in the Prevention of Atherosclerosis Trial (EPAT)
44
was a randomized
double-blind, placebo-controlled trial of unopposed oral 17-estradiol in
postmenopausal women. Women receiving estrogen had reduction in progression of
subclinical atherosclerosis measured by common carotid artery intima-media thickness
39
(CIMT) compared to women receiving placebo. We now report estrogen effects on
hemostatic and fibrinolytic factors and their association with atherosclerosis
progression.
Methods
EPAT Trial Design
Participants were non-smoking postmenopausal women, 45 years and older with LDL-
cholesterol 130 mg/dL or greater and no signs of cardiovascular disease. Eligible
subjects, randomized to oral 17estradiol (1 mg/day) (n=111) or placebo (n=111), were
followed for two years. The primary trial endpoint was rate of change in CIMT
measured at baseline and every 6 months. Ninety-seven estradiol and 102 placebo
participants had at least one repeat CIMT measurement and contributed to the primary
endpoint analysis. All participants gave written informed consent. The study protocol
was approved by the University of Southern California Institutional Review Board.
Laboratory Measurements
tPA antigen, factor VII antigen, and D-dimer were measured from samples stored at -
70
o
C from baseline, 6-month, 1-year and 2-year visits. Albumin was measured at
baseline and 3, 6 and 18 months. Due to a laboratory procedure change, only albumin
values obtained before May 2, 1996 were used. PAI-1 activity and fibrinogen were
measured from samples at the trial end; samples from earlier visits were not collected in
citrated tubes required for these measures. tPA, factor VII, and D-dimer were measured
40
in EDTA-anticoagulated plasma using enzyme immunoassays from Diagnostica Stago
(Parsippany, NJ)
15,24
. PAI-1 was measured in citrate-anticoagulated plasma using an
immunofunctional method from Trinity Biotech (St. Louis, MO). Fibrinogen was
measured in citrate plasma using the kinetic method of Clauss with reagents from
Diagnostica Stago (Parsippany, NJ)
18
. The between-run coefficients of variation were:
tPA 7%, PAI-1 6%, fibrinogen 4%, factor VII 6%, D-dimer 4%.
CIMT
CIMT was measured at baseline and every 6 months using high-resolution B-mode
ultrasonography of the right common carotid artery. Imaging methods standardized the
location and distance over which CIMT was measured ensuring the same portion of the
arterial wall was measured in each image within and across participants
44
.
Statistical Analysis
Participants with baseline and at least one on-trial CIMT measurement and one
hemostatic factor measurement (n=186) were included in these analyses. Student t-tests
and chi-square tests were used to test for baseline differences by treatment group. For
variables measured multiple times, per-participant on-trial average and average change
from baseline were compared between treatment groups with Wilcoxon rank sum tests.
Estrogen benefit on CIMT progression in EPAT was limited to subjects not taking lipid-
lowering medication. We used general linear models to test each rank-transformed
mean on-trial hemostatic factor in relation to treatment group, lipid-lowering
medication, and an interaction term of these two variables. The interaction tested if
41
treatment effects on each hemostatic variable differed by use vs. non-use of lipid-
lowering medications. Spearman correlations evaluated the linear association between
absolute CIMT levels and hemostatic variables.
We tested hemostatic variable associations with the CIMT change rate, using mixed-
effects models (random coefficients for participants) that included all CIMT
determinations (dependent variable). The independent variables tPA, factor VII, D-
dimer and albumin were modeled separately. Models were adjusted for age and BMI,
known to be related to hemostasis and atherosclerosis. All subjects were non-smokers.
We did not adjust for other cardiovascular risk factors because the path of action of a
hemostatic factor could be through such a factor. We regressed CIMT on follow-up
time (years since randomization), one time-dependent hemostatic variable (centered at
its mean baseline value), hemostatic variable * follow-up time interaction, and the
adjustment variables. The regression coefficient associated with follow-up time
represents the average rate CIMT change rate for a participant with an average baseline
value of the hemostatic factor. The regression coefficient associated with the
interaction term represents the average CIMT change rate per one-unit increase in the
hemostatic factor and tests whether the CIMT progression rate is associated with the
hemostatic factor.
Associations between each hemostatic factor and CIMT progression could result from
the hemostatic factor’s effect on atherosclerosis progression, or from an atherosclerosis
effect on the hemostatic factor. Thus, we also modeled tPA, factor VII, D-dimer, and
42
albumin as dependent variables in mixed-effect models with CIMT as the time-
dependent independent variable. Reported p-values are 2-sided; data were analyzed
using SAS 9.0.
Results
The sample was a well-educated multiethnic group with mean age 61.4 years (Table 5).
Treatment groups did not differ on baseline variables.
Table 5. Baseline Characteristics of EPAT Participants with Hemostatic Measures
Placebo (n=94) Estradiol (n=92) p-value *
Age (years) 62.1 (7.0) † 60.9 (6.6) 0.21
BMI (kg/m
2
) 29.0 (5.5) 28.9 (5.7) 0.90
Total Cholesterol (mmol/L) 6.43 (0.87) 6.54 (0.82) 0.34
HDL-Cholesterol (mmol/L) 1.11 (0.31) 1.12 (0.31) 0.79
LDL-Cholesterol (mmol/L) 3.58 (0.76) 3.71 (0.71) 0.23
Triglycerides (mmol/L) 1.83 (0.93) 1.79 (0.73) 0.75
Blood Pressure (mm Hg) 129/77 (14/7) 128/78 (14/8) 0.51/0.43
CIMT (mm) .780 (.152) .748 (.110) 0.11
Ethnicity
White 58 (61.7) 52 (56.5) 0.60
African American 8 ( 8.5) 12 (13.0)
Latina 18 (19.2) 20 (21.7)
Asian 10 (10.6) 7 ( 7.6)
Other 0 ( 0) 1 ( 1.1)
Education
High school or less 19 (20.2) 20 (21.7) 0.64
Some college 41 (43.6) 41 (44.6)
College graduate 34 (36.2) 31 (33.7)
Smoking Status
Never 52 (55.3) 43 (46.7) 0.24
Former 42 (44.7) 49 (53.3)
____________________________________________________________________
* P-value from student t-test (continuous variables) or chi-square test (categorical
variables).
† Mean (SD) for continuous variables; n (%) for categorical variables.
43
Mean pill compliance was 95% (estradiol) and 92% (placebo) (p=0.08)
44
. Serum
estradiol levels increased in the estradiol group (13.5 pg/mL at baseline, 59.6 pg/mL on-
trial; p<.001) and did not change in the placebo group (13.4 pg/mL at baseline, 14.2
pg/mL on-trial; p>0.2).
Hemostatic Factors by Treatment Group
Baseline hemostatic factors were equivalent by treatment group (Table 6). tPA, factor
VII, and albumin at each on-trial visit, average on-trial and average changes from
baseline, significantly differed by treatment. Compared to placebo, tPA and albumin
were reduced and factor VII was increased with estradiol treatment. D-dimer was not
affected by estradiol.
Figures 2-5 show tPA, factor VII, D-dimer and albumin by treatment. Mean PAI-1 and
fibrinogen at end-of-trial were significantly lower in the estradiol compared to placebo
group. Interactions between treatment and lipid-lowering medication were not
significant for any hemostatic factor (data not shown), indicating estrogen’s effect did
not differ by lipid-lowering medication use. Hemostatic factors did not differ with
lipid-lowering medication use in the total sample or within either treatment group (data
not shown).
44
Table 6. Baseline and mean on-trial values of hemostatic variables
Placebo ____Estradiol_____
N Mean (SD) N Mean (SD) p-value*
tPA (ng/mL) 86 88
Baseline 13.2 (4.5) 12.5 (4.4) 0.34
On-trial avg 13.2 (5.0) 11.1 (3.6) 0.0014
Change 0.0 (2.5) -1.5 (2.8) 0.0002
Factor VII (%) 86 88
Baseline 132.7 (31.7) 138.2 (28.4) 0.23
On-trial avg 129.2 (27.7) 148.1 (34.8) 0.0001
Change -3.5 (15.9) 9.9 (18.0)
<0.0001
D-dimer (ng/mL) 85 87
Baseline 656.2 (387.8) 595.5 (304.3) 0.25
On-trial avg 670.3 (297.3) 633.4 (304.4) 0.42
Change 14.1 (224.9) 37.9 (152.7) 0.42
Albumin (g/L) 59 61
Baseline 44.6 (2.1) 44.6 (1.8) 0.85
On-trial avg 45.3 (2.0) 43.7 (2.0)
<0.0001
Change 0.6 (1.9) -0.9 (2.1)
<0.0001
PAI-1 (U/mL)
(2 years) 72 12.4 (11.4) 67 6.6 (6.8) 0.0005
Fibrinogen (g/L)
(2 years) 71 4.07 (0.77) 67 3.64 (0.63) 0.0005
______________________________________________________________________
* p-value for difference between treatment groups calculated by Wilcoxon rank sum
test.
tPA and D-dimer were significantly correlated before (r=.21, p=0.01 at baseline; r=.20,
p=0.01 on-trial), but not after adjusting for age and BMI. PAI-1 and fibrinogen at end-
of-trial correlated with end-of-trial values as follows: PAI-1 and tPA (r=.65, p<0.0001),
45
Figure 2. Change in tPA over Time by Treatment Group.
tPA
0.00
2.00
4.00
6.00
8.00
10.00
12.00
14.00
16.00
Baseline 6 Months 1 Year 2 Years
Duration
ng/mL
Placebo
Estradiol
Figure 3. Change in Factor VII over Time by Treatment Group.
Factor VII
115.00
120.00
125.00
130.00
135.00
140.00
145.00
150.00
155.00
Baseline 6 Months 1 Year 2 Years
Duration
Percent
Placebo
Estradiol
46
Figure 4. Change in Factor VII over Time by Treatment Group.
D-dimer
0
100
200
300
400
500
600
700
800
Baseline 6 Months 1 Year 2 Years
Duration
ng/mL
Placebo
Estradiol
Figure 5. Change in Factor VII over Time by Treatment Group.
Albumin
4.200
4.250
4.300
4.350
4.400
4.450
4.500
4.550
4.600
Baseline 3 months 6 months
Duration
g/dL
Placebo
Estradiol
47
tPA and fibrinogen (r=.23, p=0.01), D-dimer and fibrinogen (r=.26, p=0.002), and PAI-
1 and fibrinogen (r=.18, p=0.03). The only correlation that remained statistically
significant after adjusting for age and BMI was PAI-1 and tPA (r=.60, p<0.0001).
Hemostatic Factor Associations with CIMT and CIMT Progression Rates
tPA was significantly positively correlated with CIMT (r=0.18, p=0.02 at baseline;
r=0.27, p=0.0004 mean on-trial values). Adjusting for age and BMI attenuated these
relationships (baseline r=0.05, p=0.53; mean on-trial r=0.13, p=0.08). Figure 6 shows
the age- and BMI-adjusted mean on-trial CIMT by quartiles of mean on-trial tPA. No
other hemostatic factor was correlated with CIMT.
Figure 6. Least squares means of on-trial CIMT adjusted for age and BMI by quartiles
of mean on-trial tPA (ng/mL).
0.71
0.72
0.73
0.74
0.75
0.76
0.77
0.78
0.79
CIMT (mm)
< 8.73
8.73 - 11.46
11.47 - 14.32
>14.32
On-trial tPA (ng/mL)
48
tPA, factor VII, D-dimer and albumin were unrelated to CIMT progression rate in the
total sample (Table 7) and when stratified by use of lipid-lowering medication (data not
shown). Absolute CIMT values were unrelated to change rates in factor VII, D-dimer
and albumin (Table 8). CIMT was significantly related to tPA change rate, with tPA
increasing faster in patients with higher CIMT (p=0.04). Stratified analysis showed the
relationship was stronger among patients not using lipid-lowering medications (no lipid-
lowering medications: beta=2.411, p=0.06; lipid-lowering medications: beta=1.180,
p=0.38) (data not shown).
Table 7. Rate of change in CIMT by hemostatic factors - estradiol and placebo groups
combined * _______________
CIMT progression rate
(µm/year) attributable
Longitudinal Effects to each effect SE p-value
Follow-up time (µm/year) -8.53† 1.987
<0.0001
Follow-up time * tPA (µm/year per ng/mL) 0.066 0.246 0.79
Follow-up time (µm/year) -8.39 1.998
<0.0001
Follow-up time * Factor VII (µm/year per %) -0.03 0.038 0.44
Follow-up time (µm/year) -8.49 2.024
<0.0001
Follow-up time * D-dimer (µm/year per ng/mL) 0.001 0.003 0.69
Follow-up time (µm/year) -4.78 2.797 0.09
Follow-up time * Albumin (µm/year per g/dL) 1.764 1.187 0.38
______________________________________________________________________
* Adjusted for age and BMI
† Average increase in CIMT progression rate for age, BMI, and applicable hemostatic
factor centered at their mean baseline values, i.e., age=61.5, BMI=29.0, tPA=12.8,
factor VII=135.4, D-dimer=628.1, and albumin=45.0
49
Table 8. Rate of change in hemostatic factors by CIMT with estradiol and placebo
groups combined * _______________
Progression rate of
hemostatic factor per
Longitudinal Effects mm CIMT (SE) p-value
Dependent variable = tPA
Follow-up time (µm/year) -0.179† 0.136 0.19
Follow-up time * CIMT (ng/mL per year per mm) 1.906 0.917 0.04
Dependent variable = Factor VII
Follow-up time (µm/year) -0.338 0.856 0.69
Follow-up time * CIMT (%/year per mm) 1.378 5.482 0.80
Dependent variable = D-dimer
Follow-up time (µm/year) 17.87 9.900 0.07
Follow-up time * CIMT (ng/mL per year per mm) 37.46 66.64 0.57
Dependent variable = Albumin
Follow-up time (µm/year) -0.005 0.017 0.91
Follow-up time * CIMT (g/L per year per mm) 0.70 3.86 0.89
______________________________________________________________________
* Adjusted for age and BMI
† Average increase in hemostatic factor for age, BMI, and CIMT centered at their mean
baseline values, i.e., age=61.5, BMI=29.0, and CIMT=0.764
Discussion
Postmenopausal estrogen therapy is associated with reduced PAI-1
19
, tPA antigen
56,81
,
and fibrinogen
19
; our results concur. Although Brown found no association between
tPA and postmenopausal estrogen therapy, these results were probably due to small
sample size. Since premenopausal levels of these factors are lower than
postmenopausal levels, these findings are not surprising
59
. Consistent with the premise
that PAI-1 and tPA levels are negatively associated with estrogen, women have lower
PAI-1 and tPA than men
28
. However, women tend to have higher fibrinogen than men
50
25,103
. Postmenopausal estrogen therapy is also associated with elevated factor VII,
confirmed in our study
19
. However, factor VII is lower in premenopausal than in
postmenopausal women and lower in women than in men
92
. As in our study, others
19,56,81
did not detect an association between postmenopausal estrogen and D-dimer.
EPAT is the first study relating reduced albumin to estrogen therapy. In summary,
postmenopausal oral 17-estradiol has favorable effects on PAI-1, tPA, and fibrinogen,
negative effects on factor VII and albumin, and no effect on D-dimer. In EPAT, the
effects of estrogen therapy on the coagulation and fibrinolytic systems appear to balance
and have no detectable effect on progression of subclinical atherosclerosis.
D-dimer, a fibrin degradation product, represents the final effect of coagulation and
fibrinolysis. Lack of an estrogen effect on D-dimer suggests that regardless of
estrogen’s effect on components of the hemostatic system, the total system remains in
balance. This argues against estrogen-related alterations in hemostasis being on the
causal pathway of atherosclerosis. Alternatively, there is greater within-subject
variation in D-dimer than other hemostatic variables. Our failure to detect a
relationship between estrogen and D-dimer may thus have been a result of insufficient
statistical power.
Although tPA antigen was significantly associated with baseline and on-trial CIMT, no
hemostatic factor was significantly associated with CIMT progression. Again, this
suggests that estrogen-related changes in hemostatic factors are not causally related to
atherosclerosis.
51
Fibrinogen, factor VII, tPA, PAI-1 and low albumin are associated with cardiovascular
disease risk, but no guidelines have been established relating specific levels to disease
risk. Meade
70
found significant positive associations of fibrinogen and factor VII with
ischaemic heart disease; levels differed in cases and controls by .41 g/L for fibrinogen
(approximately the reduction seen with estrogen in our study) and 4.1% for factor VII
(considerably less than the increase seen with estrogen in our study). Salomaa
90
found
that patients with atherosclerosis (measured by CIMT) had mean PAI-1 and tPA levels
1.4 ng/mL (1.04 U/mL) and 1.3 ng/mL, respectively, greater than those without
atherosclerosis. These differences are less than the reductions in PAI-1 and tPA seen
with estrogen treatment in this study. In a meta-analysis, Danesh
20
found a 4 g/L
decrease in albumin (considerably larger than the reduction seen with estrogen
treatment in this study) was associated with increased coronary heart disease. Estrogen
treatment thus may effect clinically meaningful changes in hemostatic factors. Our null
findings regarding CIMT progression may imply that the disease risk effects are long-
term and could not be observed in this 2-year study.
In EPAT, glucose and insulin decreased with estradiol treatment, but these decreases
were unrelated to CIMT progression
49
. Glucose and insulin levels were positively
associated with tPA (glucose r=.29, p<.0001; insulin r=.38, p<.0001) and PAI-1
(glucose r=.32, p<.0001; insulin r=.46, p<.0001), but were unrelated to the other
hemostatic measures and albumin. In EPAT, estradiol treatment did not influence blood
pressure in the total sample. However estradiol compared with placebo increased
52
systolic blood pressure (SBP) in younger women and decreased SBP in older women
100
. Changes in SBP in estradiol-treated women were positively associated with CIMT
progression
100
. Overall, systolic and diastolic blood pressure were not associated with
any measured hemostatic factor, but SBP was positively associated with tPA only
among postmenopausal women age 65 and older (r=.28, p=.04).
The null association of CIMT progression with hemostatic factors does not contradict
known associations of tPA, PAI-1, factor VII, fibrinogen, D-dimer, and albumin with
cardiovascular disease. The relationship of these hemostatic factors to other processes,
such as plaque stability, may underlie associations between hemostasis and
cardiovascular disease risk.
The positive cross-sectional association between CIMT and tPA suggests reduced
fibrinolytic activity in subjects with greater atherosclerosis, and may reflect factors
affecting both fibrinolysis and atherosclerosis including neurohormonal regulation
(affecting vasoconstriction and blood pressure), metabolic control, and inflammation.
Fibrinogen is the most studied hemostatic factor and has been related to carotid
atherosclerosis
3,10,16,33,51,58,91,113
. All but one of these studies showed a positive
relationship in at least one subgroup
10
. Three studies included men only
3,58,91
. The
study that did not show a relationship between carotid atherosclerosis and fibrinogen
included only women
10
. Since fibrinogen is higher in women than men
16,113
, its
association with atherosclerosis may vary by gender. One small study (n=38) related
53
ultrasonographically-measured carotid atherosclerosis progression (defined by a
specified percent stenosis increase) to higher fibrinogen levels
33
. Percent stenosis
measures the extent of intrusive plaque seen in more advanced atherosclerotic disease.
In the ARIC cohort, tPA, PAI-1 and D-dimer were positively related to CIMT in cross-
sectional analyses (n=914)
90
. These relationships were evident only in whites, and
when adjusted for age and multiple cardiovascular risk factors, the relationships were
no longer significant among men. Significant associations remained for tPA and PAI-1
among white women (the largest ethnic group in EPAT). In one Japanese community-
based study , PAI-1 antigen was significantly increased among men with greater CIMT;
the relationship was not evident in women
89
. In another study, PAI-1 activity was not
related to maximal CIMT
3
. No studies have reported the relationship between tPA,
PAI-1 or D-dimer and CIMT progression. The associations of hemostatic factors with
atherosclerosis are thus not consistent, possibly due to varying populations or
covariates.
This study is the first to relate rate of change in hemostatic factors to degree of
atherosclerosis (CIMT). Atherosclerotic lesions contain a lipid core with a fibrous cap
releasing pro-coagulation tissue factor when ruptured. Chronic subclinical disruption of
lesions could lead to a hemostatic system shift toward coagulation. Since coronary
artery disease is correlated with greater CIMT
65
, our finding that the rate of increase in
tPA is significantly related to CIMT supports this hypothesis. Although cross-sectional
evidence links atherosclerosis to coagulation factors, it is unlikely that changes in the
54
factors cause atherosclerosis, but more likely that progressing atherosclerosis alters the
hemostatic/fibrinolytic balance.
This is the first study to relate factor VII or albumin to ultrasound measures of carotid
atherosclerosis. It is one of few randomized clinical trials testing the association of
postmenopausal oral estrogen use with hemostatic factors and is the only study to our
knowledge relating estrogen use to albumin levels. Low serum albumin is associated
with cardiovascular disease and all-cause mortality, but the mechanisms are not well
understood
20,22
. Production of albumin declines with infection or inflammation,
suggesting albumin may be a marker for inflammation-associated cardiovascular
disease rather than a cause. However, adjusted for cardiovascular risk factors, albumin
remains associated with cardiovascular disease risk
22
. Possible causal mechanisms are
the antioxidant property of albumin
34
and inhibition of endothelial apoptosis
119
. Since
albumin binds and transports many molecules, especially fatty acids, its level may be
important to cardiovascular disease risk. Binding of sex steroid hormones may also
make albumin a regulator of the effects of hormone therapy.
We evaluated oral 17-estradiol; these results do not necessarily apply to other forms of
estrogen. The hepatic first-pass effect may be important in determining the hemostatic
effect of estrogen treatments. Post found significant decreases over one year with oral
17-estradiol in levels of fibrinogen, factor VII, tPA and PAI-1; treatment with
transdermal 17-estradiol only reduced PAI-1 levels significantly, and even these
reductions were significantly less than those seen with oral treatment
79
. D-dimer
55
changes also differed by treatment type with slight increases with oral and slight
decreases with transdermal.
Longitudinal measures of four hemostatic factors allowed us to study their relation to
atherosclerosis progression. A limitation of this study is that PAI-1 and fibrinogen were
only measured at end-of-trial. Although PAI-1 is a measure of fibrinolytic inhibition,
we did not have a corresponding measure of fibrinolytic potential provided by tPA
activity. We measured tPA antigen and found it to be highly correlated with PAI-1.
This study had a relatively small sample size and short follow-up period which may
have obscured weak associations.
In conclusion, this study confirms associations between postmenopausal oral estrogen
and increases in factor VII and decreases in tPA, PAI-1 and fibrinogen, and the null
association with D-dimer. We established an inverse association between
postmenopausal oral estrogen and albumin. Although tPA, PAI-1, D-dimer, factor VII
and fibrinogen are associated with cardiovascular disease, we found no association of
these factors to atherosclerotic progression in postmenopausal women. Nevertheless,
these results do not preclude atherosclerosis having an effect on these factors (reverse
causality) and consequently fibrinolytic activity, nor do they rule out a potential role of
these factors in atherosclerosis-associated thrombosis. The strong association of
albumin with postmenopausal estrogen highlights the need to include this variable in
future studies.
56
.
CHAPTER 3
SECOND DATA ANALYSIS
LEVELS OF CIRCULATING HEMOSTATIC FACTORS ARE RELATED TO
POSTMENOPAUSAL ESTROGEN THERAPY, BUT NOT TO CIRCULATING
ESTROGEN LEVELS
Introduction
Cardiovascular disease risk cannot be completely explained by standard risk factors.
Novel risk factors, including hemostatic factors, may help to fill in the gaps in our
knowledge and lead to novel therapies to reduce cardiovascular disease risk.
Estrogen therapy (ET) in postmenopausal women effects circulating levels of numerous
sex steroid hormones and related transport proteins. In addition, ET is associated with
changes in the levels of various coagulation and fibrinolytic factors
20-22, 24, 33, 37, 74, 81-85
. However, little work has been done evaluating the relationships
between circulating hormone levels and hemostatic factors
72,99
.
The Estrogen in the Prevention of Atherosclerosis Trial (EPAT) was a randomized,
57
double-blind, placebo-controlled trial that randomized 222 healthy postmenopausal
women to receive either unopposed oral 17-estradiol (1 mg daily) or placebo
44
.
Women receiving estrogen had a reduction in the progression of subclinical
atherosclerosis measured by rate of change in common carotid artery intima-media
thickness (CIMT) compared to women receiving placebo.
The objective of the present study was to investigate the relationship in EPAT
participants between serum levels of sex steroid hormones and transport proteins and
hemostatic factors, including tissue plasminogen activator (tPA), plasminogen activator
inhibitor 1 (PAI-1), factor VII, D-dimer, fibrinogen, and albumin.
Methods
EPAT Trial Design.
The EPAT design has been described in detail
44
. Participants were postmenopausal
women age 45 years and older with LDL cholesterol levels of 130 mg/dL or greater.
Women were excluded if they had been diagnosed with breast or gynecologic cancer in
the past 5 years or during screening, if they had previously used hormone replacement
therapy for more than 10 years or within 1 month of the first screening visit, if they had
five or more hot flushes daily that interfered with daily activity, diastolic blood pressure
greater than 110 mm Hg, untreated thyroid disease, life-threatening disease with a
survival prognosis of less than 5 years, total triglyceride level of 400 mg/dL or greater,
HDL cholesterol level less than 30 mg/dL, serum creatinine concentration greater than
58
2.5 mg/dL, or fasting blood glucose level 200 mg/dL or greater, or if they were current
smokers. Eligible subjects were randomized to oral 17estradiol (1 mg daily)
(n=111) or matching placebo (n=111), and were followed for two years. The primary
trial endpoint was the rate of change in CIMT measured at baseline and every 6 months
during the trial. All participants provided written informed consent, and the study
protocol was approved by the University of Southern California Institutional Review
Board.
Study Population.
Forty-three of the women in the EPAT parent study had no on-trial hormone data.
Therefore, 179 women were included in this sub-study.
Laboratory Measurements.
tPA antigen, factor VII antigen, D-dimer, all sex steroid hormones, and SHBG were
measured from samples stored at -70
o
C at baseline and at 6 months, 1 year and 2 years
during the trial. Albumin was measured at baseline and on-trial at 3, 6 and 18 months.
Due to a change in laboratory procedures precluding comparison of early and later
measurements of albumin, only albumin values obtained prior to May 2, 1996 were
used in these analyses. PAI-1 and fibrinogen were measured only at the end of the two-
year intervention period.
tPA antigen, factor VII antigen, and D-dimer were measured in EDTA anticoagulated
plasma using enzyme immunoassays from Diagnostica Stago (Parsippany, NJ)
15,24
.
59
PAI-1 activity was measured in citrate anticoagulated plasma using an
immunofunctional method from Trinity Biotech (St. Louis, MO). Fibrinogen was
measured in citrate plasma using the kinetic method of Clauss with reagents from
Diagnostica Stago (Parsippany, NJ)
18
.
Serum levels of androstenedione, DHEA, testosterone, estrone and estradiol were
quantified by validated, previously described RIAs
30,31
. Prior to RIA, steroids were
extracted from serum with hexane:ethyl acetate (3:2). Androstenedione, DHEA, and
testosterone were then separated by Celite column partition chromatography using
increasing concentrations of toluene in trimethylpentane. Estrone and estradiol were
separated in a similar fashion by use of ethyl acetate in trimethylpentane. SHBG was
quantified by direct immunoassays using the Immulite analyzer (Diagnostic Products
Corporation, Inglewood, CA). Free testosterone was calculated using total testosterone
and SHBG concentrations, and an assumed constant for albumin in a validated
algorithm
98,106
. Free estradiol was calculated in a similar manner.
All the hormone immunoassay methods were shown to be reliable. Specificity was
achieved by use of highly specific antisera and/or use of organic solvent extraction and
chromatographic steps prior to quantification of the analytes. Assay accuracy was
established by demonstrating parallelism between measured concentrations of a serially
diluted analyte in serum with the corresponding standard curve. Intraassay and
interassay coefficients of variation ranged from 4 to 8% and 8 to 13%, respectively. All
assay methods were found to be sensitive. The sensitivity of an RIA method was
60
determined by the smallest amount of analyte that reduced the number of counts per
minute of the radiolabeled analyte at zero mass by 2 standard deviations.
Statistical Analysis.
Student t-tests were used to test for baseline differences in continuous variables, and
chi-square tests were used to test for differences in categorical variables by treatment
group. For each participant, an on-trial average for each of the hormones was
calculated, as was the change from baseline to on-trial average. Wilcoxon rank sum
tests were used to test for differences by treatment groups in baseline, on-trial average,
and change from baseline to on-trial hormone levels. General linear models were used
to relate mean on-trial levels of tPA, factor VII, D-dimer, and albumin (dependent
variables) to mean on-trial levels of each of the hormone levels adjusted for age and
BMI, in the total sample, and stratified by treatment group. General linear models were
also used to relate end-of-trial levels of PAI-1 and fibrinogen to end-of-trial levels of
each of the hormones adjusted for age and BMI, in the total sample, and stratified by
treatment group. Additional analyses adjusted for use of lipid-lowering medication, but
since this adjustment did not materially change results these results are not shown. The
relationships between hemostatic factors and androgens and SHBG were further
adjusted for age, BMI, and estradiol level. All data were analyzed using SAS, version
9.0.
61
Results
Baseline characteristics of the women included in this study are shown in Table 9. 61%
of the study population was white. The mean age was 61.5 years and mean BMI was
29.0 kg/m
2
. The treatment groups did not differ significantly on any baseline
characteristic.
Table 9. Baseline Characteristics of EPAT Participants with Hormone and Hemostatic
Measures
Placebo (n=89) Estradiol (n=90) p-value *
Age (years) 62.2 (7.1) † 60.7 (6.6) .13
BMI (kg/M
2
) 28.9 (5.4) 29.0 (5.7) .92
Ethnicity
White 57 (64.0) 52 (57.8) .62
African American 7 ( 7.9) 11 (12.2)
Latina 16 (18.0) 19 (21.1)
Asian 9 (10.1) 7 ( 7.8)
Other 0 ( 0) 1 ( 1.1)
Education
High school or less 17 (19.1) 18 (20.0) .73
Some college 40 (44.9) 41 (45.6)
College graduate 32 (36.0) 31 (34.4)
Smoking status
Never 49 (55.1) 44 (48.9) .41
Former 40 (44.9) 46 (51.1)
Systolic blood pressure (mm Hg) 128.5 (14.2) 127.6 (14.1) .67
Diastolic blood pressure (mm Hg) 77.1 (7.0) 78.1 (7.9) .41
Total cholesterol (mg/dL) 247.0 (31.0) 252.0 (31.1) .28
HDL (mg/dL) 42.6 (11.4) 43.2 (11.9) .72
LDL (mg/dL) 138.2 (28.9) 142.5 (26.7) .30
Triglycerides (mg/dL) 160.5 (80.2) 159.7 (65.6) .94
______________________________________________________________________
* P-value from student t-test (continuous variables) or chi-square test (categorical
variables).
† Mean (SD) for continuous variables; n (%) for categorical variables.
62
Effect of Estradiol Treatment on Hormone Levels.
Women randomized to the estradiol treatment group had significantly higher on-trial
estrone, estradiol, free estradiol, and SHBG levels, and greater increases from baseline
values for these hormone measures, compared to women randomized to placebo (all
p<.0001) (Table 10).
Table 10. Baseline and mean on-trial levels of hormone and hemostatic variables
Placebo Estradiol
N Mean (sd) N Mean (sd) p-value*
Estrone (pg/mL)
Baseline 89 39.8 (13.3) 90 47.1 (31.3) .04
On-trial avg 89 48.6 (42.5) 90 315.2 (170.5) <.0001
Change 89 8.8 (42.4) 90 268.1 (169.4) <.0001
Estradiol (pg/mL)
Baseline 89 19.1 (5.3) 89 21.5 (19.8) .26
On-trial avg 89 21.2 (10.3) 89 68.9 (28.1) <.0001
Change 89 2.1 (9.4) 89 47.4 (28.3) <.0001
Free Estradiol (pg/mL)
Baseline 89 0.53 (0.18) 90 0.60 (0.52) .24
On-trial avg 89 0.58 (0.25) 90 1.62 (0.67) <.0001
Change 89 0.05 (0.19) 90 1.02 (0.70) <.0001
Testosterone (ng/dL)
Baseline 89 22.0 (9.1) 90 21.4 (10.4) .68
On-trial avg 89 22.3 (9.2) 90 22.7 (10.8) .77
Change 89 0.3 (3.6) 90 1.3 (6.2) .17
Free Testosterone (pg/mL)
Baseline 89 4.0 (1.7) 90 4.0 (2.1) .89
On-trial avg 89 4.1 (1.8) 90 3.1 (1.6)
<.0001
Change 89 0.1 (.9) 90 -0.9 (1.4) <.0001
Androstenedione (pg/mL)
Baseline 89 534.7 (227.8) 90 547.2 (248.9) .73
On-trial avg 89 520.5 (191.9) 90 495.0 (178.8) .36
Change 89 -14.2 (167.8) 90 -52.2 (175.2) .14
63
Table 10 (continued). Baseline and mean on-trial levels of hormone and hemostatic
variables
Placebo Estradiol
N Mean (sd) N Mean (sd) p-value*
DHEA (ng/mL)
Baseline 89 2.4 (1.6) 90 2.2 (1.4) .57
On-trial avg 89 2.1 (1.3) 90 1.8 (0.8) .03
Change 89 -0.2 (1.0) 90 -0.5 (1.0) .13
SHBG (nmol/L)
Baseline 89 35.2 (14.5) 89 35.0 (18.6) .94
On-trial avg 89 36.3 (17.6) 89 58.3 (23.7) <.0001
Change 89 1.0 (12.3) 89 23.3 (17.8) <.0001
TPA (ng/mL)
Baseline 84 13.11 (4.57) 85 12.59 (4.41) .45
On-trial avg 87 13.18 (5.01) 90 11.10 (3.57) .0017
Change 83 0.02 (2.49) 85 -1.52 (2.79) .0002
Factor VII (%)
Baseline 84 132.21 (30.54) 85 138.07 (27.43) .19
On-trial avg 87 128.08 (25.81) 90 149.56 (34.53) <.0001
Change 83 -3.98 (15.82) 85 9.89 (18.30) <.0001
D-dimer (ng/mL)
Baseline 83 640.53 (328.78) 84 595.32 (309.36) .36
On-trial avg 88 658.90 (272.20) 90 632.05 (303.70) .54
Change 82 22.18 (213.00) 84 35.88 (154.47) .64
Albumin (g/dL)
Baseline 76 4.5 (0.22) 77 4.49 (0.21) .57
On-trial avg 88 4.49 (0.19) 90 4.32 (0.18) <.0001
Change 76 -0.02 (0.22) 77 -0.17 (0.22) <.0001
PAI-1 (U/mL)
End of trial 69 12.58 (11.53) 66 6.71 (6.81) .0005
Fibrinogen (mg/dL)
End of trial 68 408.44 (73.85) 66 363.20 (63.66) .0002
______________________________________________________________________
* p-value for difference between treatment groups calculated by the Wilcoxon rank
sum test.
64
Treatment groups did not differ on testosterone or androstenedione levels or changes
from baseline for these androgens; however, free testosterone on-trial and change values
were significantly lower in the estradiol group than in the placebo group. The on-trial
levels of DHEA were significantly lower in the treatment than in the placebo group
(p=.03).
Relationship of Hemostatic Factors to Serum Hormone Levels.
All association statistics correlate mean on-trial levels of tPA, factor VII, D-dimer and
albumin and end-of-trial levels for PAI-1 and fibrinogen with the corresponding levels
of the hormone variables.
Estrogens
In the combined sample adjusting for age and BMI, tPA (p<.05), PAI-1 (p<.01),
fibrinogen (p<.01) and albumin (p<.01) were inversely associated and factor VII was
positively associated (p<.0001) with all the estrogens (Tables 11 and 12). The
hemostatic factors were not associated with estrogens analyzed in the placebo group
only. Factor VII was positively associated with serum estrogens in the estradiol
treatment group (p<.01 for estrone and estradiol, and p=.07 for free estradiol).
65
Table 11. Association of Mean On-trial Hemostatic Factors with Mean On-trial Hormone Levels adjusted for age and BMI
Total Sample . . Placebo . . Estradiol . p-value for treatment
Estimate (SE) p-value Estimate (SE) p-value Estimate (SE) p-value group interaction
tPA (ng/dL)
Estrone (pg/mL) -0.0041 (0.0016) .01 -0.0027 (0.0153) .86 -0.0009 (0.0019) .66 .88
Estradiol (pg/mL) -0.0256 (0.0094) .01 -0.0187 (0.0541) .73 -0.0075 (0.0121) .54 .79
Free estradiol (pg/mL) -0.8359 (0.4239) .05 2.5966 (2.5306) .31 0.0326 (0.5075) .95 .31
Testosterone (ng/dL) -0.0135 (0.0303) .66 0.0401 (0.0557) .47 -0.0465 (0.0297) .12 .15
Free testosterone (pg/mL) 0.5388 (0.1707) .002 0.8421 (0.2782) .0033 -0.0782 (0.2102) .71 .01
Androstenedione (pg/mL) 0.0013 (0.0017) .41 0.0044 (0.0027) .11 -0.0027 (0.0018) .14 .03
DHEA (ng/dL) 0.0930 (0.2955) .75 0.1569 (0.4289) .72 -0.6824 (0.4233) .11 .23
SHBG (nmol/L) -0.0621 (0.0124) <.0001 -0.1070 (0.0274) .0002 0.0280 (0.0141) .05 .01
Factor VII (%)
Estrone (pg/mL) 0.0740 (0.0121) <.0001 -0.0558 (0.0848) .51 0.0665 (0.0205) .002 .17
Estradiol (pg/mL) 0.4057 (0.0710) <.0001 -0.1713 (0.2997) .57 0.3727 (0.1289) .005 .14
Free estradiol (pg/mL) 15.7439(3.2315) <.0001 -5.0974 (14.2199) .72 10.2163(5.5658) .07 .49
Testosterone (ng/dL) 0.2908 (0.2434) .23 0.1915 (0.3095) .54 0.3426 (0.3348) .31 .66
Free testosterone (pg/mL) -1.9511 (1.4164) .17 1.1553 (1.6463) .48 -1.0349 (2.3492) .66 .40
Androstenedione (pg/mL) 0.0101 (0.0133) .45 0.0005 (0.0152) .97 0.0277 (0.0202) .17 .39
DHEA (ng/dL) -0.2208 (2.3819) .93 -0.2899 (2.3844) .90 5.9540 (4.7605) .21 .42
SHBG (nmol/L) 0.4126 (0.1025) <.0001 -0.0792 (0.1671) .64 0.3835 (0.1562) .02 .04
D-dimer (ng/dL)
Estrone (pg/mL) -0.1475 (0.1112) .19 -0.7375 (0.7439) .32 -0.2517 (0.1916) .19 .85
Estradiol (pg/mL) -1.0954 (0.6431) .09 -2.5038 (2.6292) .34 -2.1659 (1.1778) .07 .82
Free estradiol (pg/mL) -41.0258(28.7288) .16 -139.348(123.522) .26 -69.1872(49.8853) .17 .90
Testosterone (ng/dL) 0.9836 (2.0471) .63 2.0510 (2.7192) .45 -0.3189 (2.9941) .92 .44
Free testosterone (pg/mL) 11.4957(11.7669) .33 11.2785 (14.2135) .43 4.2899 (20.9024) .84 .71
Androstenedione (pg/mL) -0.0152 (0.1118) .89 -0.0026 (0.1336) .98 0.0195 (0.1816) .91 .77
DHEA (ng/dL) -24.3770(19.8226) .22 -5.6588 (20.8754 .79 -50.5266 (42.3553) .24 .62
SHBG (nmol/L) -0.9923 (0.8895) .27 -0.1939 (1.4565) .89 -1.6550 (1.4229) .25 .54
66
Table 11 (continued). Association of Mean On-trial Hemostatic Factors with Mean On-trial Hormone Levels adjusted for age and
BMI
Albumin (g/dL)
Estrone (pg/mL) -0.0003 (0.0001) <.0001 0.0000 (0.0005) .94 0.0000 (0.0001) .93 .90
Estradiol (pg/mL) -0.0020 (0.0005) <.0001 0.0005 (0.0021) .80 0.0003 (0.0007) .66 .78
Free estradiol (pg/mL) -0.0726 (0.0211) .001 0.0500 (0.0891) .58 0.0378 (0.0298) .21 .72
Testosterone (ng/dL) -0.0031 (0.0015) .04 0.0010 (0.0023) .65 -0.0042 (0.0017) .02 .22
Free testosterone (pg/mL) 0.0068 (0.0089) .44 -0.0052 (0.0119) .66 -0.0157 (0.0124) .21 .47
Androstenedione (pg/mL) -0.0000 (0.0001) .68 0.0000 (0.0001) .78 0.0002 (0.0001) .08 .16
DHEA (ng/dL) 0.0120 (0.0147) .42 0.0049 (0.0170) .78 -0.0273 (0.0253) .28 .37
SHBG (nmol/L) -0.0026 (0.0006) <.0001 -0.0005 (0.0012) .66 -0.0011 (0.0008) .20 .81
67
Table 12. Association of End-of-Trial Hemostatic Factors with Mean On-trial Hormone Levels adjusted for age and BMI
Total Sample . . Placebo . . Estradiol . p-value for treatment
Estimate (SE) p-value Estimate (SE) p-value Estimate (SE) p-value group interaction
PAI-1 (U/mL)
Estrone (pg/mL) -0.0147 (0.0048) .003 -0.0316 (0.0393) .42 -0.0039 (0.0053) .46 .40
Estradiol (pg/mL) -0.0968 (0.0286) .001 -0.0683 (0.1366) .62 -0.0496 (0.0338) .15 .90
Free estradiol (pg/mL) -3.1073 (1.2240) .01 6.5797 (6.4965) .31 -0.9212 (1.2912) .48 .17
Testosterone (ng/dL) 0.0026 (0.0900) .98 0.1858 (0.1547) .23 -0.1739 (0.0808) .04 .04
Free testosterone (pg/mL) 1.7859 (0.4599) .0002 2.6043 (0.7093) .0005 -0.7391 (0.5919) .22 .0009
Androstenedione (pg/mL) 0.0086 (0.0043) .05 0.0189 (0.0068) .0074 -0.0075 (0.0043) .09 .0016
DHEA (ng/dL) 1.9943 (0.7872) .01 2.8232 (1.1619) .02 -1.5581 (1.0427) .14 .02
SHBG (nmol/L) -0.1803 (0.0335) <.0001 -0.2780 (0.0716) .0002 -0.0796 (0.0365) .03 .0086
Fibrinogen (mg/dL)
Estrone (pg/mL) -0.1107 (0.0342) .002 0.0703 (0.2518) .78 -0.0702 (0.0461) .13 .69
Estradiol (pg/mL) -0.6738 (0.2022) .001 0.2966 (0.8723) .74 -0.5326 (0.2958) .08 .37
Free estradiol (pg/mL) -26.7962(8.6208) .002 7.2207 (41.8019) .86 -13.1751(11.3503) .25 .49
Testosterone (ng/dL) -0.9205 (0.6360) .15 -0.7500 (0.9936) .45 -1.0979 (0.7276) .14 .96
Free testosterone (pg/mL) -0.1958 (3.4593) .96 -5.7003 (4.9210) .25 -2.5333 (5.2933) .63 .83
Androstenedione (pg/mL) -0.0227 (0.0309) .46 -0.0160 (0.0460) .73 -0.0862 (0.0376) .03 .16
DHEA (ng/dL) -3.0710 (5.7868) .60 -7.2251 (7.7669) .36 -18.8502 (9.0812) .04 .14
SHBG (nmol/L) -0.4036 (0.2619) .13 0.4008 (0.5045) .43 -0.2581 (0.3346) .44 .55
68
Androgens
In the total sample, tPA and PAI-1 were positively associated with free testosterone
(p<.01 for both). These associations were entirely due to positive relationships in the
placebo group, with non-significant negative relationships in the estradiol group (p for
interaction <.01).
PAI-1 was also positively associated with androstenedione and DHEA (p<.05 for both)
due to relationships seen only in the placebo group (p for interaction <.02).. D-dimer
was not associated with any of the hormone variables (p>.09 for all).
SHBG
tPA, PAI-1, and albumin were significantly inversely associated with SHBG while
factor VII was significantly positively associated with SHBG (p<.0001 for all).
Fibrinogen was not significantly associated with SHBG. The negative associations of
tPA and PAI-1 with SHBG were much more pronounced in the placebo than in the
active treatment group (p for interaction <.01). The positive association of factor VII
with SHBG was only found in the active treatment group (p for interaction =.04).
Adjusting for serum estradiol level
Relationships between the androgen hormones and SHBG and the hemostatic factors
are shown in Table 13 adjusted for age, BMI and estradiol level. Free testosterone was
positively associated with tPA and PAI-1, and SHBG inversely associated with tPA and
69
PAI-1 (p<.01 for all). Androstenedione and DHEA were positively associated with
PAI-1 only. These results are similar to what was found before adjustment for estradiol
level with the exception that factor VII was significantly positively related to SHBG
before but not after adjustment for estradiol level.
Table 13. Association of Mean On-trial Hemostatic Factors with Mean On-trial
Androgens and SHBG Levels Adjusted for Estradiol Levels.*
Estimate (SE) p-value
tPA
Testosterone -0.0085 (0.0301) .78
Free testosterone 0.4470 (0.1759) .01
Androstenedione 0.0015 (0.0016) .38
DHEA 0.0311 (0.2937) .92
SHBG -0.0612 (0.1484) <.0001
Factor VII
Testosterone 0.2154 (0.2265) .34
Free testosterone 0.0184 (1.3611) .99
Androstenedione 0.0093 (0.0124) .45
DHEA 0.7787 (2.2142) .73
SHBG 0.1541 (0.1174) .19
D-dimer
Testosterone 1.2830 (2.0563) .53
Free testosterone 6.6856 (12.1973) .58
Androstenedione -0.0068 (0.1129) .95
DHEA -26.6915(19.9116) .18
SHBG -0.2279 (1.0637) .83
Albumin
Testosterone -0.0027 (0.0015) .07
Free testosterone 0.0029 (0.0088) .74
Androstenedione -0.0000 (0.0001) .75
DHEA 0.0086 (0.0142) .54
SHBG -0.0015 (0.0008) .04
PAI-1
Testosterone 0.0093 (0.0876) .92
Free testosterone 1.4460 (0.4729) .003
Androstenedione 0.0084 (0.0042) .05
DHEA 1.7850 (0.7716) .02
SHBG -0.1607 (0.0387) <.0001
70
Discussion
Summary of Findings.
As expected, postmenopausal oral 17-estradiol treatment was significantly associated
with higher serum levels of estrogens and SHBG and were also inversely associated
with levels of free testosterone. Although tPA, albumin, PAI-1, and fibrinogen levels
were inversely associated with estrogen levels, these relationships were not evident
when stratified by treatment group. The positive relationship between factor VII and
estrogen levels was due to a significant association in the active treatment group. Free
testosterone was positively associated and SHBG was inversely associated with tPA and
PAI-1 levels after adjustment for estradiol level. Androstenedione and DHEA were
positively associated with PAI-1 after adjustment for estradiol level. D-dimer was not
significantly related to any of the hormones in total or when stratified by treatment
group.
Comparison with Other Studies.
Postmenopausal use of estrogen has previously been shown to be negatively associated
with tPA
28,56,66,77,81,107
, PAI-1
12,19,53,107
, and fibrinogen
5,6,19,73,107
, and positively
associated with factor VII
19,73,107
. Consistent with this, we found tPA, PAI-1 and
fibrinogen were inversely associated and factor VII was positively associated with
serum estrogen levels. As in this study, a null association between postmenopausal
estrogen use and D-dimer has been reported in previous studies
19,56,81
. EPAT is the only
study that related albumin levels to postmenopausal estrogen use.
71
In a study of 42 postmenopausal women (19 current-users and 23 never-users of
hormone therapy), adjusting for age, BMI and duration of menopause, serum estradiol
levels were not related to fibrinogen, factor VII, tPA, PAI-1 or D-dimer
72
. In contrast,
we found significant (p<.01) associations in a much larger sample between estradiol and
all of these hemostatic factors with the exception of D-dimer. However, our results
substantially differ by treatment group, and we would not expect results similar to those
obtained in a study combining different proportions of current- and never-users of
hormone therapy in an observational setting as opposed to the controlled clinical trial
design of EPAT.
A second observational study relating circulating hormone levels to hemostatic factors,
the Study of Women’s Health Across the Nation (SWAN)
99
, enrolled relatively young
women (ages 42-52 years), both pre- and perimenopausal, not using hormone
replacement therapy. This large study (n=3200) provided unadjusted analyses as well
as limited analyses adjusted for BMI, study site, race, site*race interaction, time of day
of blood draw and seasonality. There was no adjustment for age or smoking, since
these variables were not associated with the study exposures or outcomes which
included levels of serum estradiol, testosterone, SHBG, fibrinogen, factor VII, tPA,
PAI-1 and D-dimer. While both the EPAT study and SWAN found a significant
adjusted negative association between estradiol and fibrinogen, EPAT, but not SWAN,
found significant relationships between estradiol and tPA, PAI-1 and factor VII. Both
studies found SHBG to be negatively associated with tPA and PAI-1, and unrelated to
72
fibrinogen after adjustment, but only EPAT found a significant adjusted positive
association between SHBG and factor VII. Although SWAN found significant adjusted
positive relationships of testosterone with tPA and PAI-1, no such relationships were
found in EPAT. Substantial differences between the population in our study
(postmenopausal women on and off hormone therapy) and that in SWAN (pre- and
perimenopausal women not on hormone therapy) may explain these differences in
results.
Although tPA, PAI-1, albumin, and fibrinogen levels are strongly inversely related to
postmenopausal estrogen therapy in the combined EPAT sample, they are not related to
estrogen levels when stratified by treatment group. Although it could be argued that
there is not sufficient variation in serum estrogen levels within treatment groups to
detect a significant association with hemostatic factors, we do see a significant
association of free testosterone with tPA and PAI-1 levels in the placebo group, and of
SHBG with tPA and PAI-1 in both the estradiol-treated and placebo groups. In fact,
free testosterone and SHBG remain significantly related to levels of tPA and PAI-1
after adjustment for serum estradiol level, and free testosterone remains significantly
related to PAI-1 after adjustment for serum SHBG (data not shown). SHBG remains
significantly related to both tPA and PAI-1 after adjusting for free testosterone. Thus, it
appears that exogenous estrogen may not affect tPA or PAI-1 through modification of
estradiol levels, but instead through its effect on SHBG and/or free testosterone.
73
The fact that SHBG was more strongly related to levels of tPA and PAI-1 in the placebo
group and to factor VII in the active treatment group may be due to threshold effects
due to the varying affinity of SHBG for different molecules. SHBG has a higher
affinity for testosterone than estrogen, and the percentage of each hormone bound
changes as the level of SHBG changes
21,87
. If the binding of SHBG to a molecule alters
its ability to affect a hemostatic factor, then SHBG would affect the hemostatic factor
differently depending on how much of it binds to the given molecule. Since how much
of it binds to the given molecule depends on serum SHBG level which differs
significantly by treatment group, the affect of SHBG on the hemostatic factor would
differ by treatment group.
The inverse association of albumin with circulating levels of estradiol in the combined
EPAT sample remained after adjustment for SHBG (p=.01), but not after adjustment for
treatment group (data not shown). Since fibrinogen levels are not related to free
testosterone or SHBG, the effect of estrogen therapy on fibrinogen levels would have to
be through another mechanism.
In EPAT, levels of D-dimer did not vary by treatment group, by levels of any of the
hormone variables, or by levels of any of the other hemostatic factors (data not shown).
Postmenopausal estrogen therapy decreases levels of fibrinogen so that less fibrin may
be produced, and since PAI-1 levels are also decreased, what fibrin is produced is more
subject to fibrinolysis. Thus, the null effect of postmenopausal estrogen therapy on D-
74
dimer levels may be the result of opposing hemostatic influences, one of which would
be expected to decrease levels of D-dimer and the other to increase levels.
In summary, postmenopausal estrogen therapy increases levels of circulating estrogens,
SHBG, factor VII, and albumin and decreases circulating levels of free testosterone,
fibrinogen, tPA, and PAI-1. In the total EPAT sample, increased levels of circulating
estrogens are associated with decreased levels of fibrinogen, tPA, PAI-1 and albumin
and increased levels of factor VII. The associations of circulating estrogen with
decreases in fibrinogen, tPA, PAI-1, and albumin however are treatment group effects;
there are no significant relationships between circulating estrogen and fibrinogen, tPA,
PAI-1, or albumin within either treatment group. The positive relationship between
circulating estrogen and factor VII is found only in the estradiol treatment group.
Similar significant associations of SHBG and free testosterone with hemostatic factors
are seen in the total EPAT sample but are not found in either treatment group. The
relationship of SHBG with factor VII, tPA and PAI-1 is statistically different by
treatment group, and the relationship of free testosterone with tPA and PAI-1 is
statistically different by treatment group.
While clinical trials have reported on relationships between postmenopausal hormone
treatment and hemostatic factors, previous studies relating circulating hormone levels in
postmenopausal women to hemostatic factors have been observational. These results
from EPAT indicate that the relationship between hormones and hemostatic factors is
complex and that carefully designed trials that eliminate many confounding factors and
75
measure the levels of circulating hormones are necessary. While postmenopausal
estrogen therapy clearly affects hemostatic variables, the mechanisms are unclear and
the implications on cardiovascular risk are uncertain.
76
CHAPTER 4
GRANT PROPOSAL
HEMOSTATIC SUPPLEMENT TO THE ELITE TRIAL
Format: NIH RO1 Grant Application
Using PHS 398 Research Plan
Specific Aims
Overall Objectives
A number of cardiovascular risk factors have been identified, leading to an easily
measured and clinically definable high-risk profile for cardiovascular events.
Nevertheless, the risk of cardiovascular disease (CVD) cannot be completely explained
by these traditional risk factors, and novel risk factors, including hemostatic factors,
may provide additional prediction of CVD risk and lead to novel therapies to reduce
CVD risk.
Postmenopausal hormone therapy (HT) is commonly seen as being pro-thrombotic,
primarily due to its association with venous thromboembolism
64
. While HT alters
77
circulating levels of many hemostatic factors, these changes are often anti-thrombotic or
fibrinolytic, leaving the relationship between HT, hemostatic factors and CVD
ambiguous. We hypothesize that anti-thrombotic and fibrinolytic alterations in
hemostatic factors may reduce the development of atherosclerosis.
Postmenopausal hormone therapy (HT) alters circulating levels of many hemostatic
factors
20-22, 24, 33, 37, 74, 81-85
. HT, however, effects levels of SHBG and sex steroid
hormones other than estradiol, yet evidence for the relationship between circulating
levels of SHBG, sex steroid hormones and hemostatic factors is very limited
108
.
Hemostatic factors have been associated with cardiovascular disease
6-12, 24-29, 35, 36
. Not
all of these relationships, however, are consistent. Whether or not changes in
hemostatic factor levels are associated with atherosclerosis is even less clear since their
relationship with cardiovascular events may be due to another mechanism such as
through thrombic events induced by fibrinolysis.
The inconsistency of studies associating postmenopausal HT with cardiovascular
disease may in part be due to a differential effect of estrogen on atherosclerosis
depending on time since menopause or extent of existing vascular disease
85
.
Furthermore, the effect of estrogen on hemostatic factors or the effect of hemostatic
factors on cardiovascular events or atherosclerosis may also differ by time since
menopause or extent of vascular disease.
78
The Early versus Late Intervention Trial with Estrogen (ELITE) is a double-blind,
placebo-controlled trial testing the effect of postmenopausal 17- estradiol on
subclinical atherosclerosis in two groups of women, those within six years of
menopause and those at 10 or more years since menopause. The primary endpoint in
ELITE is progression of subclinical atherosclerosis determined by changes in carotid
artery intima-media thickness (CIMT) measured by B-mode ultrasonography.
There is a need for randomized clinical trial data, which eliminates numerous
confounders, to evaluate the relationships between HT, circulating hormone and
hemostatic factor levels, and atherosclerosis. By using existing data and measuring
hormone levels and hemostatic factors in blood samples stored as part of ELITE, an
efficient examination of the effect of duration of menopause and atherosclerosis
severity on hemostatic factors in relation to atherosclerosis progression will be
accomplished.
Specific Aims
In a single-site, randomized, placebo-controlled, double-blind clinical trial of
healthy post-menopausal women, specific aims are:
1. To determine pro- and anti-thrombotic effects of HT as seen in levels of
hemostatic factors including factor VII, fibrinogen, von Willebrand factor
(vWF), soluble fibrin monomer (SFM), tissue plasminogen activator (tPA)
activity, tPA antigen, plasminogen activator inhibitor type 1 (PAI-1) activity,
79
and global fibrinolytic capacity (GFC). We will further test whether levels of
these factors are differentially affected by HT according to early (within 6 years
of menopause at randomization) versus late (10 or more years past menopause)
use of HT.
2. To quantify relationships between each of the above hemostatic factors and
circulating levels of sex steroid hormones, including estradiol, free estradiol,
estrone, testosterone, free testosterone, androstenedione, DHEA, and SHBG:
a. in the total sample,
b. stratified by HT vs. placebo treatment group, which will exhibit widely
different levels of circulating hormones, and
c. stratified by time since menopause group,
3. To evaluate if circulating levels of these hemostatic factors are related to the
progression of subclinical atherosclerosis measured by B-mode ultrasonography
of carotid artery intima-media thickness
a. in the total trial population, and
b. stratified by time since menopause group.
We will have longitudinal measures of all hemostatic factors and analyses will
use these factors as time-dependent variables to model the progression rate of
subclinical atherosclerosis over the trial period.
80
Background and Significance
ELITE Trial
The Early versus Late Intervention Trial with Estrogen (ELITE) (R01 AG024154,
Howard Hodis, PI) is an ongoing randomized, double-blind, placebo-controlled, long-
term (2 to 5 years), noninvasive arterial imaging trial of HT versus placebo. Active
treatment in ELITE is with 17-estradiol 1mg/day; women with a uterus also receive
4% vaginal progesterone gel (or matching placebo), 1 application every day for 10 days of
each month. The progesterone insert is an easy, practical way to deliver a progestational
agent with little systemic absorption. At the same time, it provides local delivery of the
drug to minimize endometrial hyperplasia and uterine bleeding and to produce regular
uterine bleeding patterns when bleeding occurs.
The trial uses a 2x2 factorial design with randomization on treatment assignment
(estradiol versus placebo). The second, non-randomization factor is a 2-level definition of
time since menopause. A total of 504 healthy postmenopausal women without clinical
evidence of coronary heart disease (CHD) will be randomized to 17-estradiol 1 mg/day
(with or without local progesterone, depending on hysterectomy status) versus placebo
according to their number of years since menopause, <6 years or >10 years. Recruitment
will occur over a three-year period, and the treatment duration will be between 2 and 4.5
years depending on when an individual is randomized. Candidates are telephone
81
prescreened, seen for one screening visit to determine study eligibility, and then
randomized. Participants are followed monthly for the first 6 months and then every other
month for the remainder of the trial. Rate of change in distal common carotid artery far
wall intima-media thickness (CIMT) in computer image processed B-mode
ultrasonograms is the primary trial end point.
Randomization to the ELITE trial began in Spring 2005. Serum samples are stored for
circulating hormones and all hemostatic factors proposed in this supplement are now
acquired at baseline and every 6 months of follow-up for all randomized subjects.
Rationale for the Parent ELITE Trial
More than 40 million American women are currently postmenopausal, and as the
population ages the number of women entering menopause will steadily increase. CHD
is the leading cause of mortality in women, and approximately 90% of myocardial
infarction deaths in women occur after the menopause. Although evidence is
controversial, the menopausal transition may represent a transition to increased risk of
atherosclerosis for women (see below), and the early postmenopause may offer a
limited window of opportunity for reducing this risk.
HT has serious recognized risks, including elevation in risks of breast cancer,
endometrial cancer, ovarian cancer, and thromboembolism. A number of observational
studies associate HT with reduced CHD risk, but results from several randomized
clinical trials have shown harm or an overall null effect. The Women’s Health Initiative
82
primary prevention trial
67
suggested increased CHD risk for estrogen plus progestin
(adjusted hazard ratio [HR] = 1.24, 95% confidence interval [CI] = 0.97-1.60).
However, year-by-year figures showed a hazard ratio of 1.81 for estrogen plus progestin
in year 1, decreasing to a hazard ratio of 0.70 for years six and beyond (p for trend =
.02). In addition, hazard ratios substantially differed according to years since
menopause (HR=0.89 for <10 years, HR=1.22 for 10-19 years, and HR=1.71 for >20
years) although the p-value for interaction was not statistically significant (p=.33). The
Women’s Health Initiative trial of estrogen in postmenopausal women with
hysterectomy
4
showed no effect on CHD outcomes for estrogen alone (HR = 0.91,
nominal 95% CI = 0.75-1.12). Although results by years since menopause were not
presented, results by age group (presumably well correlated with years since
menopause) showed a non-significant trend toward increased hazard ratios at higher
ages (HR=0.56 for ages 50-59, HR=0.92 for ages 60-69, and HR=1.04 for ages 70-79, p
for trend=.14).
Although there is no clear clinical trial evidence that HT reduces CHD incidence, our
group demonstrated that unopposed estradiol administered for two years reduces the
progression of subclinical atherosclerosis in women without CHD (Estrogen in the
Prevention of Atherosclerosis Trial [EPAT]
44
— see Section C.1.1). In EPAT, the
reduction in progression of atherosclerosis with estrogen therapy was most effective in
women within six years of menopause. In contrast, we and others have reported clinical
trial evidence that estradiol alone or with progestin has no effect on progression of
coronary atherosclerosis in women with established CHD
41,43
.
83
These findings gave rise to the hypothesis that HT may be cardioprotective in the early
but not late postmenopause. Because most women initiate HT around the time of
menopause
11
, most HT use in observational studies reflects use by relatively younger
postmenopausal women.
Coronary atherosclerosis begins as a fatty streak during childhood, adolescence or early
adulthood. During the decade from 45 to 55 years, women show progressive
development of raised atherosclerotic lesions, and by age 65 years women begin to
develop complicated atherosclerotic lesions that lead to plaque destabilization and
rupture
101
. In a mouse model of atherosclerosis, estrogen prevented formation of new
atherosclerotic lesions but failed to reduce established lesions
86
. In a nonhuman primate
model, immediate estrogen treatment compared to placebo reduced atherosclerotic
plaque area by 70% in animals given an atherogenic diet after ovariectomy
17
, but
estrogen had no effect on plaque area when initiation was delayed for two years after
ovariectomy
114
.
Within the vascular wall, both endothelial cells and smooth muscle cells express alpha
and beta estrogen receptors. In premenopausal women, estrogen receptor alpha
expression is reduced in atherosclerotic coronary arteries compared to
nonatherosclerotic vessels
62
. In a laboratory model, estrogen protects against
atherosclerosis when the endothelium is intact. However, if the endothelium is injured,
estrogen actually enhances cholesterol deposition
45
. In human studies, favorable effects
84
of estrogen on endothelial-mediated blood flow are seen in younger postmenopausal
women but not in older women
40
.
One inference from these studies is that HT may protect against atherosclerosis when
estrogen receptors are fully expressed in the vascular wall, when the vascular
endothelium is healthy, and when there are no complicated atherosclerotic plaques. In
other settings, however, HT may be deleterious. There is a need for clinical trials to
determine if HT mediated changes in fibrinolytic factors may be important in keeping
healthy arteries free from fibrin deposits that contribute to atherosclerotic plaque, while
fibrinolytic activity in atherosclerotic arteries may predispose them to thrombic events.
The condition of the vascular endothelium (including its efficiency in regulation of
inflammation, leukocyte and platelet adhesion, and vasodilation) may be dependent on
estrogen receptors and may also be important in determining how the hemostatic system
responds to HT. In particular, since the fibrinolytic factor tPA and, to a lesser extent, its
inhibitor, PAI-1, are produced in the vascular endothelium, the health of the
endothelium may affect production and balance of tPA and PAI-1.
Given these considerations, the primary objective in the ELITE trial is to test the
modifying effects of time since menopause on HT and progression of subclinical
atherosclerosis in healthy postmenopausal women. The hypothesis is that HT will
reduce the progression of atherosclerosis if initiated in the early postmenopause (when
the vascular endothelium is most likely to be relatively healthy) compared to the late
85
postmenopause (when the endothelium may have lost much of its responsiveness to
estrogen).
Hemostatic Factors
The hemostatic system involves a large number of variables that regulate coagulation in
the circulatory system, and others that regulate fibrinolysis. The mechanism is complex
and a balance of complimentary variables is very important to appropriate bleeding,
thrombosis and tissue repair. Hemostatic factors, including both coagulation and
fibrinolytic factors, have been recently evaluated in relation to cardiovascular risk. In
particular, fibrinogen as a risk factor for cardiovascular events has been studied at least
since the 1950s
61
. The identification of a hemostatic factor as a risk factor for
cardiovascular events, however, does not detail the mechanism of the association.
While it is possible that the hemostatic variable may act on the initiation and
progression of atherosclerosis, its effects may also involve other factors predisposing to
cardiovascular events, such as the composition of the atherosclerotic plaque and its
vulnerability to rupture. The individuals at greatest risk for progression of
atherosclerosis are not necessarily the same as those at risk for plaque rupture.
Determining the suitability of HT for individual women may be improved by
determining whether and the mechanisms by which HT-mediated changes in hemostatic
factors affect atherosclerosis and CVD risk.
Below is a brief description of each of the hemostatic factors we plan to study as part of
this supplement.
86
Factor VII (or stable factor or proconvertin) is part of the extrinsic coagulation system
which converts prothrombin to thrombin by activating factor X, ultimately leads to the
conversion of fibrinogen to fibrin
25
. Factor VII is not activated in the absence of tissue
factor, so high values of factor VII alone do not necessarily lead to coagulation.
Elevated factor VII, however, may be a marker for increased thrombic activity and has
been linked to increased risk for cardiovascular events
70
. Factor VII levels increase
with postmenopausal HT
19,73
. However, levels of factor VII in pre- and perimenopausal
women (with physiologic estrogen levels) have been found to be unrelated to circulating
estradiol levels
99
. This suggests a treatment effect on levels of factor VII that is
unrelated to circulating estrogen levels.
Fibrinogen (or Factor I) is a plasma protein produced in the liver that is the precursor
of fibrin in the coagulation process. It is higher in women than men, especially in
postmenopausal women, higher in smokers than nonsmokers, and increases with age
and body mass index
25,103
. Thrombin is required for the conversion of fibrinogen to
fibrin, so that high levels of fibrinogen alone do not necessarily lead to coagulation. In
addition, low levels of circulating fibrinogen do not necessarily indicate reduced
coagulation potential since hepatic fibrinogen synthesis can increase many times over in
acute-phase response
13
. As with factor VII, however, elevated fibrinogen may be a
marker for increased coagulation activity. Fibrinogen has been found in atherosclerotic
lesions
103
and high plasma levels have been linked with cardiovascular events in
numerous studies
2,20,47,48,68,70,112
. Fibrinogen is reduced with postmenopausal HT
5,19
.
87
Von Willebrand Factor (vWF) is produced and stored in the endothelium
60
. It binds
factor VIII and is involved in platelet adhesion at the site of endothelial injury. It
increases with endothelial injury and has been proposed as a marker of endothelial
dysfunction. The vascular endothelium expresses estrogen receptors
62,105
and these
receptors may be both important to endothelial function (including nitric oxide
production, inhibition of inflammation and leukocyte adhesion)
105
and reduced in
number and nitric oxide synthase activity with time since menopause
115
. Measuring
vWF as a surrogate marker of endothelial function would allow us to analyze whether
the effect of HT on endothelial function differs by time since menopause. vWF is
positively associated with cardiovascular disease
60
and is reduced with postmenopausal
HT
104
.
Soluble Fibrin Monomer (SFM) is produced by thrombin-induced proteolysis of
fibrinogen
74
, and thus is a measure of increased fibrin generation. SFM is further down
the coagulation pathway than factor VII or fibrinogen. It is not clear whether or not it is
associated with circulating factor VII or fibrinogen levels, although each of them is
necessary for its production. The proximity of SFM to the cross-linked fibrin matrix in
the coagulation cascade makes it an interesting variable for study in relation to
atherosclerosis. SFM has been positively associated with cardiovascular disease
74
, but
its relationship to postmenopausal HT has not been investigated.
88
tPA antigen is produced by vascular endothelial cells and activates fibrinolysis
13,50
. As
a fibrinolytic factor, tPA would be expected to be cardioprotective. However, most of
the tPA antigen measured in blood is in an inactive form bound with PAI-1
14
. In fact,
tPA antigen’s high correlation with PAI-1 activity levels, makes tPA antigen levels
essentially a surrogate measure of PAI-1
83
, i.e., high values of tPA antigen are
associated with low tPA activity. Consequently, plasma levels of tPA antigen have
been positively associated with cardiovascular events
36,47,69,81,84,90,102
, but none of these
studies adjusted for PAI-1. Plasma levels of tPA activity have been negatively
associated with cardiovascular events
27,75
, and it is important to remember that its effect
on atherosclerosis is expected to be opposite that of tPA antigen. tPA antigen levels
decrease with postmenopausal hormone therapy
56,81
but it cannot be inferred that the
effect of HT would be to increase tPA activity. Although tPA antigen and PAI-1 are
highly correlated, measuring both factors will provide greater reliability and evidence
for observed results.
PAI-1, which inhibits the action of tPA, is produced by the liver, adipose tissue, and
possibly endothelium
13,71
and is present in plasma, platelets, and extracellular fluid
103
.
Plasma levels of PAI-1 increase with age and BMI
103
and have been positively
associated with cardiovascular events
26,35,47,69,90,102
. This positive association may be
due to decreased fibrinolysis caused by PAI-1 allowing fibrin deposits to remain and to
be incorporated into atherosclerotic plaque. PAI-1 levels decrease with postmenopausal
HT
12,19
.
89
Global Fibrinolytic Capacity (GFC) is a measure of the amount of D-dimer (a fibrin
degradation product
25
) generated in a plasma sample when a standardized clot is
degraded
116
. Therefore, this test provides a measurement of the sample’s fibrinolytic
potential given a controlled fibrin quantity. High values of D-dimer are ambiguous
since they can result from either high fibrin generation or from high levels of
fibrinolysis, representing either negative or positive cardiovascular effects, respectively.
GFC eliminates the ambiguity of D-dimer measures by ensuring that increased levels of
D-dimer are not the result of high fibrin production, but rather the result of effective
fibrinolysis. Although limited study has been done using this new assay, GFC has been
found to increase with postmenopausal HT
52
.
Figure 7 is a diagram indicating the effects of some of the elements of the coagulation
and fibrinolytic systems, with the shaded boxes indicating hemostatic factors that are
included in this supplement.
90
Figure 7. Coagulation and fibrinolytic systems
Postmenopausal Estrogen and Hemostatic Factors
Table 14. Known effects of postmenopausal HT
Increases Effect of increase Decreases Effect of decrease
factor VII procoagulation fibrinogen anticoagulation
tPA activity fibrinolytic tPA antigen fibrinolytic
GFC fibrinolytic PAI-1 fibrinolytic
vWF anticoagulation
__________________________________________________________________
Numerous studies have analyzed the relationship of postmenopausal HT and levels of
circulating hemostatic factors, and have found that HT decreases tPA antigen
56,81
, PAI-
1
12,19
, fibrinogen
5,19
and vWF
104
, and increases factor VII
19,73
and GFC
52
(Table 14).
None of these studies, however, have considered time since menopause, level of
binds collagen
and platelets on
vascular
endothelium
prothrombin thrombi plasmin plasminogen
fibrin standardized
clot
factor VIIIa
factor VIIa
factor VIII
bound to vWF
tissue factor
tPA
D-dimer
SFM
PAI-1
GFC
vWF
factor VII
fibrinogen
X
Coagulation Fibrinolysis
91
atherosclerosis, or age as possible modifying factors in their analyses. The effect of HT
on SFM is unknown.
The vascular benefits of estrogen are estrogen receptor mediated and depend upon the
integrity and functional status of the vascular endothelium
62,80
. Advancing age and
atherosclerosis are associated with vessel wall injury, with depletion of the numbers and
function of estrogen receptors and compromise of endothelial function. Since vWF,
tPA, and possibly a portion of PAI-1 are produced in the vascular endothelium
13,60
, the
effect of HT on levels of these factors, and on the levels of factors that are downstream
of them in the hemostatic model, may depend on age, time since menopause, and extent
of existing atherosclerosis. Table 15 summarizes hypothesized effects of
postmenopausal HT on CIMT and hemostatic factor levels in the total sample
population and by early vs. late postmenopause group.
Table 15. Hypothesized effects of postmenopausal HT
Overall Early postmenopause Late postmenopause
group group
CIMT Decrease Decrease Null
Factor VII Increase Null Increase
Fibrinogen Decrease ? ?
vWF Decrease Decrease Null
SFM ? ? ?
tPA antigen Decrease Decrease Null
tPA activity Increase Increase Null
PAI-1 Decrease Decrease Null
GFC Increase Increase Null
______________________________________________________________________
92
Associations Between Hemostatic Factors and Circulating Sex Steroid Hormones
Although HT decreases tPA antigen, PAI-1, and fibrinogen, and increases factor VII,
there is evidence that these alterations are not related to the circulating estrogen
levels
108
. The HT association with hemostatic factors may be due to the effect of HT on
other hormonal factors, particularly SHBG. In EPAT we found that higher levels of
serum estrogens were associated with decreased tPA antigen, PAI-1, and fibrinogen,
and increased factor VII
108
. However, treatment-stratified analyses showed no effect of
circulating estrogen levels on these hemostatic factors except for increased factor VII in
the active treatment group. SHBG was positively associated with factor VII in the
active treatment group and negatively associated with tPA antigen and PAI-1 in both
treatment groups.
Little other work has been done relating circulating sex steroid hormone levels to
hemostatic factors. In a study of 42 postmenopausal women (19 current-users and 23
never-users of hormone replacement therapy)
72
, with adjustment for age, BMI and
duration of menopause, serum estradiol levels were not significantly related to
fibrinogen (r=-.25, p=.21), factor VII (r=-.04, p=.83), vWF (r=.16, p=.36), tPA (r=-.11,
p=.53), or PAI-1 (r=-.13, p=.51). Not only did this study lack power due to small
sample size, but adjusting for duration of menopause in this study cannot accomplish
the same goal as stratification since we hypothesize that we will see effect modification
by time since menopause. Thus, these results are not inconsistent with our hypothesis
that HT effects factor VII, fibrinogen, tPA, and PAI-1 through some mechanism other
than estradiol level.
93
A second observational study relating circulating hormone levels to hemostatic factors
is the Study of Women’s Health Across the Nation (SWAN)
99
, which enrolled relatively
young women (ages 42-52 years), both pre- and perimenopausal, not using hormone
replacement therapy. This large study (n=3200) provided unadjusted analyses as well
as limited analyses adjusted for BMI, study site, race, site* race interaction, time of day
of blood draw and seasonality. There was no adjustment for age or smoking, since
these variables were not associated with the study exposures or outcomes which
included serum levels of estradiol, testosterone, SHBG, fibrinogen, factor VII, tPA,
PAI-1 and D-dimer. Both EPAT and SWAN found a significant adjusted negative
association between estradiol and fibrinogen. EPAT, but not SWAN, found significant
relationships between estradiol and tPA, PAI-1 and factor VII. Both studies found
SHBG to be negatively associated with tPA and PAI-1, and unrelated to fibrinogen after
adjustment, but only EPAT found a significant adjusted positive association between
SHBG and factor VII. Although SWAN found significant adjusted positive
relationships of testosterone with tPA and PAI-1, no such relationships were found in
EPAT. It is important to remember, however, that in EPAT the relationships between
hemostatic factors and hormones were frequently different in the estradiol vs. placebo
groups. Substantial differences between the population in EPAT (postmenopausal
women on and off HT) and that in SWAN (pre- and perimenopausal women not on HT)
may explain these differences in results. Clearly, the association of circulating SHBG
and sex steroid hormones with hemostatic factors remains unclear and further studies,
ideally controlled clinical trials, are needed.
94
Hemostatic Factors and Subclinical Atherosclerosis
The hemostatic factors discussed above have been associated with cardiovascular
disease
6-12, 15, 25, 27, 29, 31, 32, 35, 36
, but their relationships with atherosclerosis and
atherosclerosis progression are less clear. Most studies relating hemostatic factors with
atherosclerosis are cross-sectional, frequently with atherosclerosis assessed as a
dichotomous variable. Different methods to determine atherosclerosis are used with
inherently different study populations. Results are inconsistent and much uncertainty
remains. Hemostatic factors may contribute to the progression of atherosclerosis and/or
they may contribute to cardiovascular events by contributing to plaque instability. We
hypothesize that the effect of hemostatic factors on cardiovascular disease depends on
the condition of the vascular endothelium. A high fibrinolytic potential in a healthy
endothelium may quickly clear small thrombi before they have the chance to be
incorporated into plaque. On the other hand, high fibrinolytic potential in a vessel with
unstable plaque could lead to plaque rupture resulting in a cardiovascular event.
Therefore, an increase in fibrinolytic potential mediated by HT may slow
atherosclerosis progression in women with recent menopause with relatively healthy
vessels while leading to ischemic events in women with longer time since menopause.
Increased levels of circulating coagulation factors in response to HT may not be
harmful when the vasculature is healthy; however, when plaque ruptures, the resulting
thrombus could lead to a cardiovascular event. We have included vWF as one of the
factors we will be studying since it is known to be a marker for endothelial
dysfunction
60
, and therefore its level at baseline may be an effect modifier for the
95
association of postmenopausal HT with the level of other hemostatic factors and/or the
progression of atherosclerosis.
A large body of evidence indicates that HT affects levels of circulating hemostatic
factors and that levels of hemostatic factors are related to CVD. Furthermore there is
much evidence to suggest that the effect of HT on cardiovascular disease may vary by
time since menopause. The parent trial to this supplement will test differences in
atherosclerosis progression in response to HT in women who are in the early vs. late
postmenopausal period. This supplement would take the logical step to further
investigate HT and early vs. late menopause in relation to hemostatic factors that may
be involved in atherosclerosis progression.
We will have longitudinal measures of CIMT, circulating levels of sex steroid
hormones, and hemostatic factors. Repeated measures of variables will provide greater
power and minimize the effect of variation over time, as well as provide information on
change at an individual level, and be a better indicator of causality. Few studies of
hemostatic factors have been done with longitudinal measures and this study will allow
an important addition to the literature.
Rationale for Funding Request for a Supplement to the ELITE Trial
The relationship between hemostatic factors and the progression of atherosclerosis has
been studied very little. The design of ELITE will ensure a substantial age range in our
study population. (As of March 2007 the average age in the early menopause group
96
was 55.5 years and the average age in the late menopause group was 64.5 years.) In
EPAT we found significant positive associations of age with levels of fibrinogen, D-
dimer and tPA antigen. Therefore, in ELITE we expect to see a range of values for
these hemostatic factors, consistent with a population of healthy postmenopausal
women. The effect of hemostatic factors on atherosclerosis, as well as the effect of HT
on hemostatic factors, in both early and late menopause will be testable. ELITE
contains several exclusion criteria, including clinical signs, symptoms or history of
cardiovascular disease, and thus patients are not expected to exhibit extreme levels of
atherosclerosis. Nevertheless, this trial including early and late menopausal women will
provide a population of diverse age and extent of subclinical atherosclerosis. ELITE
provides a platform upon which effects of HT on hemostatic variables can be
determined at very modest incremental cost. Since the most costly portion of any trial
is recruitment and follow-up, adding a component to study hemostatic factors and
circulating hormone levels in this trial will be simple and will not require any design
change. For both the early and late postmenopause, the added value for studying
hemostatic factors in this estradiol treatment trial is immense, and an opportunity that
we believe should not be overlooked. There are, in fact, three important questions that
will be addressed by this supplement: 1) Does the effect of HT on levels of circulating
hemostatic factors vary by early vs. late menopause? 2) Are levels of circulating
hemostatic factors related to levels of circulating estrogens, androgens, or SHBG
independent of treatment group and might these relations differ by time since
menopause? 3) Are hemostatic factors related to the progression of subclinical
97
atherosclerosis in the total population and/or in subpopulations defined by time since
menopause or severity of atherosclerosis at baseline?
Timeliness and Uniqueness of the ELITE Supplement
As findings from the Women’s Health Initiative became widely known to women and
their physicians, the number of prescriptions for HT declined dramatically
42
. However,
nearly 10 million women continue to use HT
42
, primarily for treatment of moderate to
severe vasomotor symptoms. Thus, millions of women continue to use HT, and these
women will initiate HT almost entirely as younger women in the early postmenopause.
A satisfying answer to the question of how HT affects their cardiovascular health
cannot be obtained without an understanding of the mechanisms (hemostatic and others)
underlying the results of the various hormone studies, as well as an understanding of the
different associations that may be observed in different subpopulations. ELITE will
obtain these data on two subpopulations of primary interest in regard to HT and CVD
risk, women in the early versus late postmenopause.
Little information is available regarding the association between levels of hemostatic
factors and circulating hormonal factors. HT is known to affect levels of many
hormonal factors other than estrogens and progestins, such as SHBG and free
testosterone
108
. The availability of blood samples that will show a wide range in both
hemostatic and hormonal factor levels makes this an ideal setting for studying the
relationships between these factors in postmenopausal women with variation due to
exogenous HT. Uniquely, this study will, in a randomized, placebo-controlled trial,
98
study not just the effect of HT on hemostatic factors, but also the relationships between
hemostatic factors and atherosclerosis progression and between levels of circulating
hormones and hemostatic factors.
Significance of the ELITE Supplement
Currently, 38% of women in the United States are 45 years of age and older. By 2015,
this percentage will rise to approximately 50%
109
. With the growing number of women
entering menopause and introduction of new hormonal products into the marketplace,
the market for postmenopausal hormone therapy will increase. While findings from the
Women’s Health Initiative seem to have resulted in a rapid shift in prescription
practices regarding postmenopausal HT, not all treatment regimens were affected
equally, and approximately 57 million prescriptions were written for 10 million
postmenopausal women in 2003
42
. This indicates that women and their physicians want
the benefits of HT but are concerned about possible risks. Therefore, understanding the
effect of estrogen on the progression of subclinical atherosclerosis, especially in young
postmenopausal women who are likely to initiate these products for menopausal
symptoms continues to be an important and timely public health issue. The parent trial
to this proposed supplement will importantly address the effectiveness of HT in two
distinct subpopulations of women, but will not address specific mechanisms of any
reduction in the progression of subclinical atherosclerosis. Regardless of the result of
the parent trial, this supplement will provide information on the effect of HT on levels
of hemostatic factors and their role in the progression of subclinical atherosclerosis in
women who are in early vs. late postmenopause. A better understanding of hemostatic
99
factors in relation to HT may aid in the development of novel therapies for the reduction
in atherosclerosis progression in postmenopausal women.
Previous Work/Preliminary Studies
Investigator Experience in Other Randomized Controlled Trials in Postmenopausal
Women
Because of the dearth of information concerning atherosclerosis progression and its
determinants and treatment in women relative to men, we have directed much of our
research effort into the area of atherosclerosis imaging and women's cardiovascular health.
Preliminary data led to the Women’s Estrogen-Progestin Lipid-Lowering Hormone
Atherosclerosis Regression Trial (WELL-HART), a randomized, double-blind,
placebo-controlled, serial coronary angiographic trial of HT versus placebo in
postmenopausal women with established CHD (Section C.1.2) and the Estrogen in the
Prevention of Atherosclerosis Trial (EPAT), a randomized, double-blind, placebo-
controlled, carotid artery ultrasound IMT trial designed to test the effects of unopposed
estrogen therapy versus placebo on the progression of subclinical atherosclerosis in healthy
postmenopausal women without preexisting CHD (Section C.1.1).
Relevant data from EPAT and WELL-HART are presented to demonstrate our
experience and success in recruiting, randomizing, and conducting trials in
postmenopausal women. These two sister trials were conducted concurrently. Principal
100
results from EPAT were published in the Annals of Internal Medicine
44
and those of
WELL-HART were published in the New England Journal of Medicine
43
.
EPAT
EPAT was a randomized, double-blind, placebo-controlled, carotid artery ultrasound trial
designed to test whether unopposed micronized 17-estradiol (1 mg daily) versus placebo
reduces progression of subclinical atherosclerosis (carotid IMT) in 222 healthy
postmenopausal women without preexisting CHD with LDL-C levels of at least 130
mg/dL. All subjects were non-smokers at randomization and remained so on-trial. After
randomization, subjects were treated for 2 years with unopposed estrogen therapy or
placebo and with lipid-lowering medication (primarily HMG-CoA reductase inhibitors) if
needed, to maintain LDL-C levels <160 mg/dL. All subjects received dietary counseling
according to step II American Heart Association dietary recommendations. Carotid artery
ultrasonography was performed at baseline (2 visits) and every 6 months on-trial. The
primary trial end point was the rate of change in the distal common carotid artery far wall
IMT in computer image processed B-mode ultrasonograms. 42% of the EPAT cohort was
comprised of women from a racial minority group
44
.
As shown in Table 16, on-trial estrogen and placebo pill compliance was high, 95% and
92%, respectively. There was an appropriate rise in the mean serum estradiol level in the
estrogen-treated group and no change in the mean serum estradiol level in the placebo
group.
101
Table 16. On-trial comparison of medication compliance by treatment group in EPAT
Placebo (n=102) 17-Estradiol (n=97)
Variable Mean (SD) Mean (SD) p-value
17-estradiol compliance (%) 92.2 (11.0) 94.9 (11.1) .08
Estradiol level (pg/ml)
*Baseline 13.4 (4.8) 13.5(6.4) .90
On-trial 14.2 (10.1) 59.6 (35.1) .0001
* Subjects with baseline and at least one follow-up carotid artery IMT determination.
In EPAT the primary trial end point was the rate of change in the distal common carotid
artery far wall IMT in computer image processed B-mode ultrasonography. Table 17
summarizes EPAT results and shows that treatment with 17-Estradiol vs. placebo was
associated with a statistically significant reduction in the progression of CIMT.
Table 17. Comparison of CIMT by treatment group* in EPAT
Placebo 17 -Estradiol
Variable N Mean (SE) N Mean (SE) P-value
Overall Cohort:
Baseline CIMT (mm) 102 0.776 (0.0149) 97 0.752 (0.0111) .48
Rate of CIMT change (mm/yr) 102 0.0036 (0.0026) 97 -0.0017 (0.0019) .045
No lipid-lowering therapy 35 0.0134 (0.0047) 42 -0.0013 (0.0032) .002
* Intent to treat analysis; subjects with baseline and at least 1 on-trial CIMT measurement.
WELL-HART
WELL-HART was a randomized, double-blind, placebo-controlled, serial coronary
angiographic trial. A total of 226 postmenopausal women 48 to 76 years old (mean age of
63.5 years) with established CHD were randomized to one of three treatment arms: daily
micronized 17 -estradiol 1 mg per day plus medroxyprogesterone acetate placebo, daily
micronized 17 -estradiol 1 mg per day plus active medroxyprogesterone acetate 5 mg per
102
day for 12 days each month, and placebo. Subjects also received dietary counseling and
lipid-lowering medication (primarily HMG-CoA reductase inhibitors) to maintain LDL-C
levels <130 mg/dL. Coronary angiograms were obtained prior to randomization and
repeated three years after randomized treatment under standardized protocol. The primary
trial end point was the mean per patient change in percent diameter stenosis determined by
quantitative coronary angiography. 70% of the WELL-HART cohort was comprised of
women from a racial minority group
43
.
As Table 18 shows, on-trial 17 -estradiol, medroxyprogesterone acetate, and placebo
compliance was at least 90% in the three treatment groups. There was an appropriate rise
in the mean serum estradiol level in the hormone treated groups and no change in the mean
serum estradiol level in the placebo group.
Table 18. On-trial comparison of study medication compliance by treatment group in
WELL-HART*
Placebo 17 -Estradiol 17 -Estradiol
(n=72) (n=72) +MPA (n=72)
Variable Mean (SD) Mean (SD) Mean (SD) P-value
17 -estradiol (%) 90 (14) 90 (15) 90 (15) .90
MPA compliance (%) 93 (17) 90 (22) 90 (21) .67
Estradiol level (pg/ml)
Baseline 13.2 (7.3) 13.7 (8.4) 12.8 (5.4) .74
On-trial 13.0 (6.4) 39.5 (24.3) 43.4 (29.6) <.0001
*MPA = medroxyprogesterone acetate
In summary, we successfully recruited and randomized 448 postmenopausal women (42%
minority in EPAT, 70% minority in WELL-HART) over a 2-year period. Independently
and combined, treatment compliance in EPAT and WELL-HART was exceptional (Tables
103
16 and 18). With our safety protocol and close monitoring procedures (see Protection of
Human Subjects) there were no serious uterine events as a result of HT. The dropout rate
in EPAT and WELL-HART was excellent for long-term hormone therapy (6% per year in
WELL-HART) as was the number of participants contributing to the primary trial end
point (90% in EPAT).
Investigator Experience in Randomized Controlled Trials Measuring Hemostatic and
Hormonal Factors
All of the hormonal factors (estrone, estradiol, free estradiol, testosterone, free
testosterone, androstenedione, DHEA, and SHBG) proposed to be analyzed as part of
this study have previously been studied (collected and analyzed) as a supplement to
EPAT. tPA antigen, Factor VII, fibrinogen, and PAI-1 activity have also been studied
as a part of EPAT
107
. Results showed that postmenopausal HT was associated with
changes in levels of hemostatic factors (Table 19) and that changes in hemostatic
factors were not related to progression of subclinical atherosclerosis. The analysis of
vWF, SFM, tPA activity, and global fibrinolytic capacity will be new to this
supplement.
104
Table 19. Baseline and mean on-trial values of hemostatic variables in EPAT
Placebo ____Estradiol_____
N Mean (SD) N Mean (SD) p-value*
tPA (ng/mL) 86 88
Baseline 13.2 (4.5) 12.5 (4.4) 0.34
On-trial avg 13.2 (5.0) 11.1 (3.6) 0.0014
Change 0.0 (2.5) -1.5 (2.8) 0.0002
Factor VII (%) 86 88
Baseline 132.7 (31.7) 138.2 (28.4) 0.23
On-trial avg 129.2 (27.7) 148.1 (34.8) 0.0001
Change -3.5 (15.9) 9.9 (18.0)
<0.0001
PAI-1 (U/mL)
(2 years) 72 12.4 (11.4) 67 6.6 (6.8) 0.0005
Fibrinogen (g/L)
(2 years) 71 4.07 (0.77) 67 3.64 (0.63) 0.0005
______________________________________________________________________
* p-value for difference between treatment groups calculated by Wilcoxon rank sum
test.
Collaborative Work
Drs. Hodis and Mack have a strong history of collaborative research in atherosclerosis.
Dr. Mack (director of the Data Coordinating Center) and Dr. Hodis (director of the
Atherosclerosis Research Unit) were co-investigators on the WELL-HART, VEAPS
and EPAT trials. In addition, they are on-going collaborators in the B-VAIT, WISH
and ELITE trials. They have co-authored over 50 papers. In addition, Dr. Chandler, a
recognized expert in hemostatic factors, collaborated with Drs. Hodis and Mack in the
EPAT trial. Dr. Stanczyk, an expert in sex steroid hormones, also collaborated with
Drs. Hodis and Mack in the EPAT and WELL-HART trials, as well as collaborating
with Dr. Hodis on other smaller studies involving HT
95,96,111
.
105
ELITE Recruitment to Date
As of March, 2007, 385 women (76% of the projected total randomization of 504
women) were randomized in ELITE. This is is a well-educated group of women
including 28% of minority ethnicity. The mean age in the subgroup of women within 6
years (mean 3.3 years) of menopause is 55.5 years and the mean age in the subgroup of
women with ten or more (mean 14.8 years) years since menopause is 64.5 years.
Recruitment is expected to be complete in July, 2008.
Methods
Overall Design of the ELITE Trial
For this randomized, double-blind, placebo-controlled, noninvasive ultrasonographic
trial, 504 eligible postmenopausal women without clinical evidence of preexisting CHD
will be recruited. Menopause can be either a natural menopause or, alternatively, an
induced menopause that includes bilateral oophorectomy. (Women with hysterectomy
but without bilateral oophorectomy or unknown oophorectomy status are ineligible.) In
addition, women must be within 6 years of menopause (early postmenopause) or at least
10 years beyond the menopause (late postmenopause); women between 6 and 10 years
postmenopausal are ineligible.
Recruitment is in progress and will occur over three years, and the total treatment
period for each participant will be between 2 and 4.5 years, depending on when the
106
participant is randomized. Candidates are prescreened by telephone, seen for one
screening visit to determine eligibility, and then randomized.
The primary trial end point of the parent trial is the rate of change in the common
carotid artery IMT in computer image processed B-mode ultrasonograms, a noninvasive
measurement of subclinical atherosclerosis. Within each postmenopausal stratum (early
or late postmenopause), women are randomized into 2 arms within substrata defined by
common carotid artery IMT (<0.75 mm, >0.75 mm) and type of menopause (natural or
surgical). The primary endpoints of the proposed hemostatic supplement will be
hemostatic factor levels. Analyses will test for treatment group differences and with
circulating sex steroid hormone levels, and the progression of subclinical
atherosclerosis measured by B-mode ultrasonography. Primary analyses will be
conducted separately within each postmenopausal stratum as well as in the total sample.
Active treatment is with oral 17-estradiol 1 mg/d with identical-appearing placebo oral
preparation. Women with a uterus also receive 4% vaginal progesterone gel, 1
application every day for 10 days of each month, or an identical placebo application.
107
Inclusion Criteria:
1) Postmenopausal woman (final menstrual period at one year prior) and serum
estradiol level <20 pg/ml, who are either
a) Fewer than 6 years from menopause (early postmenopause group), or
b) At least 10 years from menopause (late postmenopause group)
Exclusion Criteria:
1) Clinical signs, symptoms, or personal history of CHD
2) Hysterectomy without oophorectomy
3) Diabetes mellitus or fasting serum glucose > 126 mg/dl
4) Uncontrolled hypertension (diastolic blood pressure >110 mmHg)
5) Thyroid disease (untreated)
6) Renal insufficiency (serum creatinine >2.0 mg/dL)
7) Life threatening illness with prognosis <5 years
8) Current use of HT
9) Current use of lipid-lowering agent
Type and Frequency of Laboratory Variables Collected in ELITE
Table 20 summarizes the schedule of the ELITE parent trial laboratory determinations.
In the parent trial, standard lipids are determined every 6 months. Direct measurement
of the lipoproteins by preparative ultracentrifugation with standardization to the CDC
using Lipid Research Clinic protocol is also performed. Plasma estradiol and
108
progesterone levels are determined every 6 months by standard radioimmunoassay
techniques. A standard chemistry panel is determined every 12 months and a complete
blood count and thyroid stimulating hormone level determined at baseline.
Table 20. Laboratory variables funded in ELITE
Variable Determination frequency
Direct lipid quantification: Baseline, every 6 months
total cholesterol, total triglyceride,
LDL-C, VLDL-C, HDL-C
LDL-TG, VLDL-TG, HDL-TG
Estradiol and progesterone Baseline, every 6 months
Chemistry panel Baseline, every 12 months
Complete blood count Baseline
Thyroid stimulating hormone Baseline
In addition, samples are being obtained and stored every six months for DNA, serum
and plasma. For this proposed supplement, additional determinations using stored
samples will include factor VII, fibrinogen, vWF, SFM, tPA antigen, tPA activity, PAI-
1, GFC, estrone, free estradiol, testosterone, free testosterone, androstenedione, DHEA and
SHBG. Table 21 shows the proposed schedule of determination for these variables.
Table 21. Laboratory variable measurements proposed for this supplement
Variable Determination frequency
Sex steroid hormones: Baseline, every 6 months
estrone, estradiol, free estradiol
testosterone, free testosterone
ndrostenedione, DHEA, SHBG
Hemostatic factors: Baseline, every 6 months
Factor VII, tPA antigen and activity,
Fibrinogen, VWF, SFM, PAI-1, GFC
109
Laboratory Methods
All hormones, SHBG and hemostatic factors will be analyzed from stored samples at
the end of the ELITE trial and run in batch to eliminate inter-batch variability.
Hemostatic Factors
Factor VII and tPA antigen are measured in EDTA anticoagulated plasma using enzyme
immunoassays from Diagnostica Stago (Parsippany, NJ)
15,24
. PAI-1 activity is
measured in citrate anticoagulated plasma using an immunofunctional method from
Trinity Biotech (St. Louis, MO). Fibrinogen is measured in citrate plasma using the
kinetic method of Clauss with reagents from Diagnostica Stago (Parsippany, NJ)
18
. In
addition to fibrinogen and PAI-1, vWF, SFM, and GFC determinations require citrated
plasma samples and tPA activity requires an acid citrate tube (trade name Stabilyte). A
test kit for SFM is currently available from Diagnostica Stago.
Sex Steroid Hormones
Serum levels of androstenedione, DHEA, testosterone, estrone and estradiol are
quantified by validated, previously described RIAs
30,31
. Prior to RIA, steroids are
extracted from serum with hexane:ethyl acetate (3:2). Androstenedione, DHEA, and
testosterone are then separated by Celite column partition chromatography using
increasing concentrations of toluene in trimethylpentane. Estrone and estradiol are
separated in a similar fashion by use of ethyl acetate in trimethylpentane. SHBG is
quantified by direct immunoassays using the Immulite analyzer (Diagnostic Products
110
Corporation, Inglewood, CA). Free testosterone is calculated using total testosterone
and SHBG concentrations, and an assumed constant for albumin in a validated
algorithm
98,106
. Free estradiol is calculated in a similar manner.
All the immunoassay methods were shown to be reliable. Specificity was achieved by
use of highly specific antisera and/or use of organic solvent extraction and
chromatographic steps prior to quantification of the analytes. Assay accuracy was
established by demonstrating parallelism between measured concentrations of a serially
diluted analyte in serum with the corresponding standard curve. Intraassay coefficients
of variation ranged from 4 to 8%. All assay methods were found to be sensitive. The
sensitivity of an RIA method was determined by the smallest amount of analyte that
reduced the number of counts per minute of the radiolabeled analyte at zero mass by 2
standard deviations.
Ancillary Data
A number of measures relevant to the analysis and interpretation of this sub-study are
collected as part of the data collection process in the main study. A number of these
measures have been shown to be related to hemostatic factors (for example, age, BMI,
smoking status). We will evaluate these variables for treatment group differences.
Variables found to differ between treatment groups and to confound the association
between hormones and hemostatic factors will be used as covariates in the analysis of
treatment group differences. These measures include:
111
1. Basic demographic data (age, ethnicity, years of education, occupation, smoking
status)
2. Reproductive history, including past use of postmenopausal HT (duration of use
and time since last use).
3. Presence and severity of menopausal symptoms (evaluated at baseline and
during the trial).
4. Blood pressure (measured at baseline and every 2 months on-trial)
5. Lipids (measured at baseline and every 6 months on-trial)
6. Common carotid artery IMT (measured at baseline and every 6 months on-trial)
7. Use of non-study medications, especially ACE inhibitors which may affect
hemostatic factors (assessed at every study visit)
8. Anthropometric factors: weight, height, BMI, waist and hip circumference,
waist-hip ratio (measured at baseline and every 6 months on-trial)
Data Coordination of Hemostatic Factor Supplement
We will use the existing resources of the Data Coordinating Center for the ELITE Trial,
directed by Dr. Mack, for data entry, management and statistical analysis.
We will use the Web-based data entry system using SQLt that has been constructed for
the parent ELITE study. Data entry involves review of study forms for accuracy and
completeness. At data entry time, certain fields are programmed for required entry and
range checks are used when applicable. Laboratory data are obtained on hard copy and
112
as electronic files sent to the Data Coordinating Center. Hard copies of the data also
follow for verification of electronic transmission. Procedures for electronic receipt of
these study data are in place and practiced from previous studies including EPAT. SAS
programs are used for verification of IDs, dates and study visit numbers, and for
identification of logical errors, missing values, and additional range checks. Audit trails
are used to monitor the status of all data queries. Routine data reports to the study
investigators and relevant clinic staff include recruitment status, participant follow-up
(participants late for visits, missed visits, dropouts, deaths, trial completers), study
medication compliance (with identification of individual participants who are poor
compliers), and adverse events.
Mirroring our existing data management and reporting programs, SAS programs will be
written to assess the quality of the hemostatic sub-study data (including verification of
IDs, dates and study visit numbers, and identification of implausible values and missing
data).
Statistical Power Considerations for the Hemostatic Factor Supplement
ELITE is a 2x2 factorial design clinical trial, with randomization on one factor (estradiol
vs. placebo). The second factor is time since menopause (<6 vs. >10 years). Sample size
requirements for the parent trial were estimated to test the hypothesis that an effect of
estradiol treatment on the carotid IMT progression rate will significantly vary by time
since menopause. The overall study hypothesis for the parent trial thus involves a test for
treatment-by-time since menopause interaction in the 2x2 factorial design. Sample size
113
estimates for the parent trial to test this interaction at a power of 80% and a 2-sided alpha
level of 0.05, accounting for anticipated dropout, was 126 in each of the four strata, for a
total of 504 subjects (252 per treatment group combining the menopause strata).

Power to Detect Differences in Levels of Hemostatic Factors by Group
Overall comparison of treatment groups (menopause strata combined): For our
sample size of 504 women (252/group), we will be able to detect treatment effect sizes
(mean group difference divided by pooled SD) of 0.264 and greater assuming a 10%
dropout rate (227/group completing) and 0.279 and greater assuming a 20% dropout
rate (202/group completing), at a 2-sided alpha level of 0.05 and 80% power.
Therefore, this study will be able to detect treatment group differences of about 1/4 of
the standard deviation of the hemostatic factor.
Treatment group comparisons within menopause strata: Within each menopause
stratum, 126 subjects per group will be randomized. We will be able to detect treatment
effect sizes of 0.374 and greater assuming a 10% dropout rate (113/group completing)
and 0.396 and greater assuming a 20% dropout rate (101/group completing), at a 2-
sided alpha level of 0.05 and 80% power. Thus, within each menopause stratum, we
will be able to detect treatment group differences of about 37-40% of the standard
deviation of the hemostatic factor. As shown in Table 22, the detectible treatment
group difference within menopause strata is smaller than the treatment group
differences observed in the hemostatic factors in EPAT. If effects on the hemostatic
114
factors that are new to this supplement are comparable to those for the factors measured
in EPAT, we have sufficient power to detect them.
Table 22. Detectible treatment group differences within menopause strata for alpha=.05
and 80% power assuming 10% dropout rate
Standard deviation Treatment group Detectible
Observed in EPAT difference observed difference for
in EPAT ELITE
supplement
Factor VII (%) 33.0 18.9 12.3
Fibrinogen (g/L) .737 .431 .276
tPA antigen (ng/mL) 4.4 2.1 1.6
PAI-1 (U/mL) 9.8 5.8 3.7
______________________________________________________ _______________ _
Power to Detect Associations between Hormonal and Hemostatic Factors
For the association between hormonal and hemostatic factors, the estimated 454
participants (504 less 10% dropout rate) will allow us to detect correlations of at least
0.13 at a power of 80% and 2-sided alpha of 0.05. Associations evaluated within
menopause strata or within treatment group (227 anticipated participants for each group)
will be able to detect correlations of 0.18 and higher. Table 23 shows the correlation
coefficients between hormones and hemostatic factors observed in EPAT. We do not
expect to find significant correlations between all hormones and hemostatic factors, but
with the power provided by this supplement, even relatively modest correlations will be
detectible.
115
Table 23. Correlation coefficients between hemostatic factors and sex steroid hormones
observed in the total EPAT sample
Factor VII Fibrinogen tPA antigen PAI-1
Estrone .34 -.19 -.14 -.24
Estradiol .32 -.19 -.13 -.22
Free estradiol .27 -.20 -.05 -.13
Testosterone .15 -.04 -.03 -.04
Free testosterone -.04 .07 .26 .29
Androstenedione .09 -.11 .02 .08
DHEA 0 -.15 -.08 .09
SHBG .29 -.16 -.41 -.53
Power to Detect Relationships Between Hemostatic Factors and Subclinical
Atherosclerosis
We will test for associations of each hemostatic factor with the progression of
subclinical atherosclerosis as determined by CIMT. This will be a test of the treatment
group difference between regression slopes obtained by modeling the progression of
subclinical atherosclerosis as a function of the hemostatic factor level in the total group
as well as in each of the time since menopause groups (<6 years vs. >10 years). With
an estimated 454 study participants (504 less 10% dropout rate), an alpha level of .05,
and 80% power, we will be able to detect a difference in the rate of change of CIMT of
0.5 µm/year per % factor VII and of 4.0 µm/year per ng/mL tPA antigen. While these
are substantially larger than the effects seen in EPAT (which showed no association),
the other primary aims of this proposal are adequately powered, and the value of
confirming EPAT findings, that levels of hemostatic factors do not affect the
progression of subclinical atherosclerosis, merits the statistical analysis.
116
In EPAT, we only had end-of-trial measurements of fibrinogen and PAI-1. While these
hemostatic factors were not related to CIMT on a cross-sectional basis, the finding of a
null relationship with CIMT progression to be tested here would indicate that a
mechanism for the relationships between these hemostatic factors and CVD should be
investigated in areas other than atherosclerosis, such as acute thrombotic events. We do
not have preliminary data for the relationship of vWF, SFM, tPA activity, or GFC with
CIMT, and this supplement will provide valuable data.
Statistical Analyses of the Hemostatic Factor Supplement
Preliminary analyses
All treatment group analyses will follow the intent to treat principle. In other words, all
participants who have applicable data will be included in the analyses, regardless of
their compliance to their randomized treatment. Following the statistical analysis plan
for the main study, we will evaluate the baseline comparability of the two treatment
groups (overall and within menopause strata) using t-tests for independent samples or
chi-square tests, depending on the type of data. Important baseline variables to be
compared between treatment groups are age, serum hormone and SHGB levels,
hemostatic factor levels, past use of menopausal hormone therapy (duration and time
since last use), presence of menopausal vasomotor symptoms, type of menopause
(natural or surgical), blood pressure and use of ACE inhibitors. Note that because we
are randomizing within the divergent time since menopause strata (<6 years and 10
years), we do not expect to achieve baseline comparability between the two duration since
117
menopause groups. However, we do expect to achieve and will evaluate baseline
comparability between treatment groups.
Analyses of Relation between Postmenopausal Estrogen Use and Hemostatic
Factors (Specific Aim 1)
The primary comparison to address the first study aim will be a comparison of treatment
differences (placebo vs. estradiol) in the total sample population and within the 2 groups
of time since menopause. All tests will use an overall 2-sided alpha of 0.05.

For each participant, a change score (mean on-trial average minus baseline score) will
be computed for each hemostatic factor. The primary analytical methods will be the
student t-test (or a non-parametric analog) for each of the hemostatic factors. All
treatment group summaries (means, SEMs, for estradiol vs. placebo) will be presented in
the total sample (irrespective of time since menopause) and also within each of the time
since menopause groups. If analyses of baseline comparability outlined above reveal
group differences on baseline variables that are related to study endpoints (hemostatic
factors), the treatment group differences will then be compared using analysis of
covariance, with the baseline variables that differ by group added as model covariates.
Analyses will not include covariates relating to on-trial levels of variables differing
between treatment groups even when those variables may be related to study endpoints.
For example, on-trial lipid levels will likely differ by treatment group, but since our
interest is in the effect of treatment and duration since menopause group on hemostatic
118
factors regardless of path, on-trial lipid levels will not be included in the models.
General linear models will use one hemostatic factor as the dependent variable, and
treatment group, menopause stratum, and a treatment group x menopause stratum as
independent variables to test for a differential effect of HT on hemostatic factors
according to early vs. late menopause. We will evaluate key assumptions of these
methods, including normality of the change scores and homogeneity of variances
between treatment groups. Transformations of the study variables or nonparametric
procedures will be used if the model assumptions are violated.
Analyses of Relation between Levels of Sex Steroid Hormones and SHBG and
Hemostatic Factors (Specific Aim 2)
The relationships between circulating levels of each of the hormonal factors and the
hemostatic factors will be analyzed. We will use general linear models to modelthe
mean on-trial value of each hemostatic factor (dependent variable) as a function of the
the mean on-trial value of each hormone and SHBG (independent variable). Age,
BMI and smoking status, which are related to both hormone and hemostatic factor
levels, will be included as covariates in the models. Additional covariates may be
included in ancillary analyses. For example, relationships between hormones and
hemostatic factors controlling for SHBG will be of interest. Summaries for the
relationships ( estimates, SEMs, and p-values) between each hormonal and each
hemostatic factor will be presented in the total sample and also within each of the
treatment group and time since menopause strata. General linear models will also be
119
used to test for a differential effect of HT with menopause stratum. The value of each
mean on-trial hemostatic factor will be modeled as a function of menopause stratum, the
value of each mean on-trial hormone/SHBG, and an interaction term of menopause
stratum x hormone/SHBG, as well as age, BMI, smoking status, and other related
covariates as determined above. The interaction term p-value will be reported.
Analyses of Relation between Hemostatic Factors and Sub-clinical Atherosclerosis
(Specific Aim 3)
We will analyze CIMT in relation to each hemostatic factor in a time-dependent
manner. The rate of change in CIMT will be computed for each subject by fitting a
regression line of CIMT on years since baseline ultrasound examination. The estimated
slope of the regression line will be used as that subject’s CIMT rate of change. To
accomplish this, we will use mixed effects models for longitudinal data. Each CIMT
measure (baseline and on-trial) will be included as a dependent variable. The regression
coefficient associated with the model covariate of time on-study will estimate the
average progression rate of CIMT. Random effects terms for the regression intercept
and slope (change over time) will allow for individual subject deviations from the
average baseline CIMT and CIMT rate of change, respectively. Analyses will be done
in the total sample and stratified by each of the study-design factors (treatment group
and time since menopause group) as well as on baseline CIMT dichotomized at the
median. The statistical test for the overall difference in the CIMT rate by level of each
hemostatic factor will be accomplished by adding a 2-way hemostatic factor by time on-
study interaction term. The significance of the regression coefficient associated with
120
this interaction term is the test that the CIMT rates differ by hemostatic factor in the
overall sample or the subgroup evaluated. If any baseline variables (lipids, blood
pressure, etc.) are found to be differentially distributed among the treatment groups at
baseline, we will perform the data analyses with and without adjustment for the relevant
baseline variable.
Study Timeline
Year 1: Develop additional computerized database, error check and quality control
programs for additional laboratory data (modification of EPAT protocol).
Years 2-3: End of ELITE trial: Laboratory analyses of stored samples. Perform data
entry and quality control checks; edit and maintain study database.
Year 3: Statistical analyses and manuscript preparation.
Pretection of Human Subjects
Since this supplement is proposed to be added to the ELITE trial, we have provided the
Protection of Human Subjects section relevant to the entire trial including the risk and
protection associated with the hemostatic component for review by the IRB.
121
Risks to Human Subjects
Human Subjects Involvement and Characteristics
Proposed involvement of human subjects in the work outlined in the research
design and methods section: First contact with individuals for this trial will occur
through their response to a mailing, advertisement, word-of-mouth, etc. when the
individual contacts the Atherosclerosis Research Clinic for further information
regarding the trial. At this prescreening stage, the individual will be asked questions to
determine her eligibility and level of interest. The first on-site contact with the
individual will occur at the screening visit where informed consent will be obtained. In
addition, a few questionnaires will be administered to further determine eligibility and
level of interest and to collect baseline data. Vital signs by standard techniques will be
determined for eligibility. Blood from the antecubital vein by standard procedures will
be collected to determine whether the potential participant meets the plasma estradiol
level defined for trial eligibility. If the individual meets the eligibility criteria and
remains willing to participate in the trial, she will return for a baseline visit at which
further questionnaires will be administered and the first of 2 baseline ultrasound scans
of the carotid arteries will be performed. In addition, a transvaginal ultrasound will be
performed in women with a uterus prior to the randomization clinic visit to determine
uterine wall thickness. A breast and pelvic examination will be performed and a Pap
smear obtained. Women without a uterus will have a breast and vaginal examination
and a Pap smear of the blind vaginal pouch will be obtained. All participants will have
a mammogram. The participant will return to the clinic in 1 to 2 weeks and a final
eligibility determined along with the results of the mammogram and breast and vaginal
122
examinations. At the randomization visit, the participant will randomly receive in a
double-blinded fashion either the active estrogen product or placebo, undergo the
second baseline ultrasound scan of the carotid arteries, will be administered
questionnaires to collect trial-related research information, have an ECG performed,
vital signs determined including weight and measurement of waist and hip
circumference and baseline blood samples collected. A cognitive test battery will also
be administered at the randomization visit. On-trial, participants will undergo carotid
ultrasound examinations every 6 months, have blood samples collected from the
antecubital vein every 6 months, be administered study questionnaires every 6 months,
have dietary counseling at every clinic visit after completing 3-day dietary diary
booklets, return flushing diaries at every visit and have an ECG, mammogram, Pap
smear and breast and vaginal examination performed annually. Women with a uterus
will have a transvaginal ultrasound performed annually and if the uterine wall is >5 mm
thick an endometrial biopsy will be obtained. At the 2 year trial visit, the cognitive test
battery will be re-administered.
Characteristics of the subject population, including anticipated number, age range,
and health status: We plan to randomize 504 healthy postmenopausal women (no
vaginal bleeding for 1 year and serum estradiol level <20 pg/ml) who are <6 years and
>10 years postmenopausal and do not have preexisting cardiovascular disease or
diabetes mellitus and are not currently using lipid-lowering medication. Women who
have had only a hysterectomy and no oophorectomy will be excluded since determining
the actual time of menopause will be difficult. By study design, half of the participants
123
will be less than 6 years since menopause when randomized and half will be 10 years or
more since menopause when randomized. Based on our previous studies of similar
design (EPAT, WISH, and BVAIT) and the same population from which the cohort for
this new proposal will be recruited, we anticipate the age range to be from 30 to 80
years old. Since the health status of the cohort is defined as healthy a priori, there will
be very few major illnesses that develop during the trial.
Identify the criteria for inclusion or exclusion of any subpopulation: There are no
pre-defined subpopulations planned for this trial. However, subjects will be
randomized according to their number of years since menopause. The
inclusion/exclusion criteria will apply to all potential participants: inclusion: 1)
postmenopausal females (no vaginal bleeding for 1 year and serum estradiol level <20
pg/ml) who are <6 years and >10 years postmenopausal, and exclusion: 1) clinical
signs, symptoms, or personal history of cardiovascular disease; 2) women with
hysterectomy only (no oophorectomy); 3) diabetes mellitus or fasting serum glucose
>126 mg/dL; 4) uncontrolled hypertension (diastolic blood pressure >110 mmHg); 5)
thyroid disease (untreated); 6) renal insufficiency (serum creatinine >2.0 mg/dL); 7) life
threatening illness with prognosis <5 years; and, 8) current use of lipid-lowering
medication.
Rationale for special classes of subjects: This trial does not include fetuses, neonates,
pregnant women, children, prisoners, institutionalized individuals, or others who may
124
be considered vulnerable populations. However, because of the scientific question, this
trial is limited to postmenopausal women.
Sources of Materials
There are 12 sources of materials to be obtained from participants in this trial: 1)
ultrasound images of the carotid arteries; 2) questionnaires to collect trial-related
research information; 3) vital signs including brachial artery blood pressure, pulse rate,
weight, and waist and hip circumferences; 4) blood specimens drawn from the
antecubital vein by standard procedures, including DNA specimens; 5) annual ECG; 6)
annual transvaginal ultrasound measurements of uterine wall thickness; 7) endometrial
biopsies when indicated; 8) annual pelvic examinations with Pap smears; 9) annual
breast examinations and mammograms; 10) dietary information obtained from 3-day
dietary diary booklets; 11) flushing information obtained from monthly diaries; and, 12)
medical records to verify clinical events. All information acquired in relation to this
trial is for research purposes only.
Potential Risks
There are two broad classes of risk to consider for this trial, those related to the
procedures of the trial and those related to the intervention under study. Although any
risk can be potentially serious, based on our experience in conducting two hormone
therapy trials in 448 women for 2 to 3 years, none related to this trial are likely to be
serious in nature. This is a non-invasive imaging end point trial with standard of
practice minimally invasive procedures (blood draws, transvaginal ultrasounds, Pap
125
smears and endometrial biopsies, when indicated) used to study an intervention with
well understood side effects.
Blood draws may lead to bruising, which could become infected. In over 10 clinical
studies conducted over 25 years by our group, other than a few ecchymoses, no serious
complications have ever occurred. Carotid ultrasonography could potentially lead to
bradycardia or stroke if there is severe bilateral carotid artery luminal obstruction and
great pressure is exerted to the carotid artery under interrogation. In over 10,000 carotid
artery ultrasound examinations conducted by our group, no complications from carotid
ultrasonography have ever occurred. Abrasions, uterine rupture and infection are all
potentially serious but very unlikely complications of transvaginal ultrasound, Pap
smears and endometrial biopsies. Vaginal discomfort from transvaginal
ultrasonography is also possible but with a thin sleek probe this is rare. However,
typically there is some discomfort with Pap smears and endometrial biopsies. In over
1,200 transvaginal ultrasounds, Pap smears and endometrial biopsies performed in our 2
hormone therapy trials, EPAT and WELL-HART, we experienced no complications.
The nature of some of the questions assessed by questionnaire is personal and may
create feelings of embarrassment or infringement of privacy. However, psychological
damage is very unlikely.
Potential risks to hormone therapy include uterine and breast cancer and
thromboembolism. However, based on our experience with 448 postmenopausal
women followed for 2 to 3 years in 2 hormone therapy trials, EPAT and WELL-HART,
126
the likelihood of these events is low with our safety protocol and close monitoring
procedures. In both these trials, EPAT and WELL-HART, there were 3 total breast
cancers, 2 placebo subjects (1 in EPAT and 1 in WELL-HART) and 1 estrogen-
progestin treated subject (in WELL-HART), 1 pulmonary embolus in an estrogen-
treated subject (in WELL-HART) and no uterine cancer associated with hormone
therapy.
In summary, although there is always risk of potentially serious complications in any
trial, the nature of the procedures and close monitoring of the study intervention make
these complications unlikely, as our experience with 2 previous hormone therapy trials
demonstrates. As outlined below, we have aggressive safety testing procedures proven
effective in our 2 previous hormone therapy trials in place.
Risks to Human Subjects Specific to this Proposal
No additional subjects will be recruited for this supplement, nor will any additional
contact with human subjects be necessary to complete the aims of this supplement.
This supplement involves analysis of blood samples obtained and stored as part of the
parent trial as well as additional statistical analyses involving data collected as part of
the parent trial. All subjects included in the parent trial are eligible for inclusion in this
supplement.
127
Adequacy of Protection Against Risks
Recruitment and Informed Consent
We will use 2 predesignated sources of recruitment for healthy individuals; The Kaiser
Permanente Health Care System and the USC Health Network. Concurrently, media
sources will be utilized; the Los Angeles Times newspaper, USC Trojan family
magazine, USC Health and Medicine magazine, USC Chronicle newspaper, and public
service announcements.
Written informed consent will be obtained at the screening visit for participation by
either Dr. Hodis or his designee. Purpose, benefits, and risks of the study will be
explained to each participant on an individual basis. All procedures and accompanying
risks will be explained in detail with time for discussions and questions. Subjects will
be given a copy of the informed consent recapitulating the verbal explanation of trial
risks and benefits. Subjects will be asked to sign the informed consent forms in the
presence of a witness.
Protection Against Risk
All risks are minimized by using highly trained, experienced, professional staff. Our
record over 25 years in conducting more than 10 clinical studies that have used both
coronary angiography and carotid ultrasonography have resulted in minimal risk,
including a hormone therapy trial with a coronary angiographic end point (WELL-
HART) and a hormone therapy trial with a carotid ultrasonography end point (EPAT).
128
We have had no complications referable to the proposed imaging end point, carotid
ultrasonography in over 10,000 procedures; this fact is a prime reason why we choose it
for the asymptomatic population under study in this trial. In the event of any life
threatening medical emergency occurring within the clinic, staff as well as physicians,
are trained in cardiopulmonary life support procedures. The Atherosclerosis Research
Clinic is situated between two large University of Southern California hospital facilities
to which access is readily available. For less than life threatening events, all subjects
will be under the care of their private physicians for all needs save the study protocol.
For the study protocol, subject safety will be monitored by clinic staff and physicians,
including well trained gynecologists with particular expertise in reproductive
endocrinology and postmenopausal health. Subjects are informed of abnormal study
results and a letter with the abnormal results is sent to the subject's personal physician
with the subject’s permission. If any laboratory result is potentially life threatening
(e.g., pancytopenia), Dr. Hodis immediately communicates with the subject's private
physician after communicating with the subject.
In terms of the study products to be used, the potential risks to hormone therapy
including uterine and breast cancer are well known. We plan to use a similar safety
protocol and close monitoring procedures for cancer surveillance in this proposal as we
have used in our 2 previous hormone therapy trials. An annual pelvic examination with
Pap smear and transvaginal ultrasound and breast examination with mammogram will
be performed at baseline and on an annual basis. Endometrial biopsies will be
performed if the endometrial thickness is >5 mm. In addition, as part of the research
129
program, subjects will be instructed on self-breast examination procedures and will be
given American Cancer Society approved instructional materials. We will be
conservative and will monitor for potential side effects with annual CBC and chemistry
panels. It should be noted that with our safety protocol and close monitoring
procedures used in EPAT and WELL-HART, we experienced no serious uterine events
as a result of hormone therapy in the 448 women followed for 2 to 3 years across 2
hormone therapy trials (see Previous Work/Study Results). Further, uterine protection
from hyperplasia will be provided with the use of locally applied vaginal progesterone
gel every month.
All information, data, and test results are treated in the strictest sense of patient-doctor
confidentiality and our clinic staff operates under the same guidelines. Subjects’ charts
are kept at the Atherosclerosis Research Center (Clinical Science Center building) in
locked offices only accessible to research staff. Hard copies of data are kept at the Data
Coordinating Center (Center for Health Professionals building) stored in locked
cabinets in private offices without public access. All subject data are given a unique
identification number and subject anonymity is maintained by using these identification
numbers without identifiers that could link the data to a subject. The computerized
databases are on a secured network that requires special administrative approval by Drs.
Mack or Azen to access the study directory. The only individuals with administrative
access are the data monitor, data manager, and the primary statistician. Passwords are
given only to individuals who are given access to the study data. The password is
regularly changed and access is closely monitored. All data are coded and subject
130
identifiers are removed. In over 25 years of conducting clinical investigation, we have
never had a breach in the confidentiality of subject data. In short, our procedures for
subject protection are time-tested across many studies, are sound and the likelihood of
our procedures being successful for this trial is very high. All Atherosclerosis Research
Center personnel are HIPPA trained and certified, and the standard operating
procedures of the Atherosclerosis Research Center follow HIPPA regulations.
Likewise, all Center personnel are IRB certified in the protection of human subjects.
Adequacy of Protection Against Risks Specific to this Proposal
As with other data collected in the parent study, the strictest confidentiality and security
procedures will be used regarding the hemostatic analyses of stored samples, and the
statistical data generated will be available only to the data monitor, data manager and
the statistician in data files containing no personal identifiers.
Potential Benefits of the Proposed Research to the Subjects and Others
Although there are several health-related benefits associated with the proposed trial,
there is no guarantee that an individual will directly benefit from them. All participants
will have intensive dietary counseling in the modification of dietary habits that could
lead to a reduction in cardiovascular disease and cancer. Health monitoring/screening
procedures may also afford benefits to participants, including cholesterol
measurements, blood pressure determinations and uterine and breast examinations. In
addition, chemistry panels, CBCs, thyroid function tests and other laboratory
measurements will be performed at baseline and/or throughout the trial. Subjects are
131
informed of abnormal study results and a letter with the abnormal results is sent to the
subject and the subject's personal physician, with the subject's permission. Hormone
therapy is known to reduce osteoporosis, bone fractures and colorectal cancer.
Additionally, the trial result itself will provide benefit to those who participated in the
study as well as to other postmenopausal women through the importance of the
knowledge to be gained, as outlined below as to whether early postmenopausal
initiation of HT can reduce the progression of atherosclerosis in healthy young
postmenopausal women.
There are no potential benefits to the subjects specific to this proposal, with the
exception of the knowledge to be gained, as outlined below.
Importance of the Knowledge to be Gained
The importance of the knowledge to be gained impacts 2 important and integrated areas,
one scientific and the other public health. From a scientific standpoint, understanding
whether hormone therapy impacts atherosclerosis progression according to the timing of
the intervention creates a whole new paradigm when considering interventions for
atherosclerosis not previously considered nor tested in human clinical trials. If this new
paradigm proves valid, then future trials will need to consider the timing of
interventions when testing the effects of potential antiatherosclerotic therapies.
Although this paradigm is clearly operative in animal models, understanding its import
in humans is imperative in order to avoid disregarding potentially effective
interventions when not proven effective in a population without consideration of timing
132
of initiation. The question as to whether HT is atheroprotective in younger versus older
postmenopausal women is not only timely, but also of immense medical, financial and
public health importance. This is particularly important since the number of women
entering menopause is rapidly increasing and other forms of therapy for menopausal
symptoms, in particular flushing are not very effective. In addition, new
postmenopausal hormone products are entering the marketplace. With all these factors
considered, postmenopausal hormone use is likely to increase, especially in the young
postmenopausal woman. Knowledge as to whether hormone therapy is effective in
reducing the progression of atherosclerosis in young postmenopausal women will assist
women in making a more informed decision when initiating postmenopausal hormone
therapy. It is important to understand whether hormone therapy has an
antiatherosclerotic effect when initiated early in the postmenopausal period so that
women can make a truly informed decision concerning their expectations of this form
of therapy. Although strongly supported by the totality of data, this question remains
unanswered in properly conducted randomized controlled trials and there are no on-
going trials to directly address this question. It is imperative that the effect of early
postmenopausal initiation of hormone therapy on atherosclerosis progression be
validated or refuted. The risks to subjects who participate in this trial are small. The
procedures used in this study are not investigational and are well-validated and safe.
All of the procedures used in this trial have been used in many previous studies by other
investigators around the world as well as by the investigators themselves with negligible
complications. The estrogen compound is commercially available. Therefore, the small
risks of this clinical trial are outweighed by the potential knowledge to be gained; that
133
is, does early initiation (defined as time since menopause) of HT reduce the progression
of atherosclerosis in healthy young postmenopausal women without preexisting
cardiovascular disease, the women most likely to initiate hormone therapy.
Importance of the Knowledge to be Gained Specific to this Proposal
The knowledge to be gained from this specific proposal cannot be obtained from the
parent trial alone, but the parent trial provides a cost-effective manner of obtaining
important additional knowledge regarding the effect of postmenopausal hormone
therapy on hemostatic factors, the association between hemostatic factors and
circulating levels of sex steroid hormones, and the relationship between hemostatic
factors and the progression of subclinical atherosclerosis. The relationship between
postmenopausal hormone therapy and hemostatic factors is just beginning to be tested.
Study results have not always been consistent and variations may be due to differences
in the study populations, including time since menopause. The relationships between
levels of hemostatic factors and levels of circulating sex steroid hormones have been
even less studied. It is not clear whether associations that have been found between
hemostatic factors and cardiovascular disease are a result of associations between
hemostatic factors and atherosclerosis, or between hemostatic factors and acute
thrombic events. In fact, the associations may be complete opposites depending on the
state of the vascular endothelium related to time since menopause. As with the parent
study, this supplement can help to define those women who would be most likely to
clinically benefit (due to hemostatic alterations linked to decreases in cardiovascular
disease risk) from postmenopausal hormone therapy. In addition, this supplement will
134
provide important knowledge regarding the relationships between hemostatic factors
and atherosclerosis that will be useful in assessing cardiovascular risks in the general
population.
Data and Safety Monitoring Plan
An External Data and Safety Monitoring Board (EDSMB) was established according to
NIH guidelines and meets on a yearly or more frequent basis if deemed necessary to
review safety data, to advise the investigators on conduct of the trial and to make
decisions concerning continuance of the trial based on all available data. The Data
Coordinating Center will prepare reports (overseen by Dr. Azen) on all safety and
compliance data and data that relate to end point analyses. The EDSMB is composed of
individuals with expertise in clinical trials, statistics, hormone therapy and
cardiovascular disease. The study reports prepared for the EDSMB list treatment
groups by letter designation (A and B) and not by therapy. Only for safety reasons and
upon request by EDSMB for the specific areas of concern will treatment group
assignments be provided to the EDSMB. This procedure has worked well for all of our
clinical trials over the last 15 years and we have never had unblinding of a study due to
premature release of information. All adverse events that occur in our trials are
reported to the University of Southern California (USC) IRB within two business days
after learning of the event in accordance with our standard operating procedures. If any
life-threatening or fatal event is deemed attributable to the intervention under study, it is
reported to the USC IRB and NIH within 24 hours after learning of the event.
135
Inclusion of Women and Minorities
Inclusion of Women
In the context of the hypothesis and scientific objectives, only postmenopausal women
will be enrolled into this study.
Inclusion of Minorities
Participants in this trial will represent the ethnic distribution of Los Angeles County
with 30% of the cohort from a racial minority group (see Targeted/Planned Enrollment
and Inclusion Enrollment Report Format Pages). The rationale for this selection is to
satisfy our objective of applicability of the trial results to the population as a whole. No
racial or ethnic group will be excluded from this trial. As of March, 2007, 28% of the
340 women recruited were minorities. We are implementing recruitment strategies
successfully used in our 2 previous randomized controlled hormone therapy
atherosclerosis trials in postmenopausal women, EPAT and WELL-HART in which
42% and 70% of the subjects, respectively were from a racial minority group. We also
have alternative and specific recruitment strategies for the proposed recruitment of 30%
minorities in the unlikely event that the current recruitment veins fail to provide
sufficient minority participants. These include the Los Angeles County Health Care
System that is comprised of approximately 90% minorities and the Los Angeles County
Employee Retirement Association which is comprised of approximately 60%
minorities. Each of these minority-rich recruitment veins has been used successfully in
our previous trials to over-sample or specifically study minority populations. For
136
example, the Los Angeles County Health Care System was used as the sole source to
recruit 95% minorities (primarily Latinos and African Americans) to our trial of
atherosclerosis and insulin resistance, the Troglitazone Atherosclerosis Regression
Trial. The cohort for this trial of 299 men and women was recruited in less than 2
years. ELITE recruitment began in the spring of 2005 and is expected to be complete in
July, 2008.
Inclusion of Children
This proposal does not include individuals under the age of 21 years since it is not
relevant to children. This is a study of postmenopausal women.
Vertebrate Animals
Vertebrate animals are not involved.
Data Sharing
The proposed research will include data from approximately 504 women recruited
predominantly from 2 sites in the Greater Los Angeles area, the University of Southern
California and the Kaiser Permanente Medical Center. While we are committed to
making the data available, we must safeguard the rights and privacy of participants.
Even though the final data set will be redacted to strip all identifiers prior to release for
sharing, the risk of deductive disclosure of participant identities remains since this is a
137
relatively small sample drawn from a restricted geographic area with linkage to other
data bases at their respective institutions. Thus, we will make the data and associated
documentation available to users under the auspices of the PI through a data-sharing
agreement that provides for: a) IRB approval for the proposed research using the data;
b) a commitment to using the data only for research purposes and only for the research
approved by IRB; c) a commitment not to identify any individual participant; d) a
commitment to securing the data using appropriate computer technology and other
necessary safeguards; e) a commitment to not transfer the data to other users; and, f) a
commitment to destroying or returning the data after analyses are completed. In
accordance with NIH policy, data sharing will occur in a timely fashion defined as no
later than the acceptance for publication of the primary findings from the final dataset.
138
Bibliography
1. Abbott RD, Petrovitch H, Rodriguez BL, Yano K, Schatz IJ, Popper JS, Masaki
KH, Ross GW, Curb JD. Ankle/brachial blood pressure in men >70 years of age
and the risk of coronary heart disease. American Journal of Cardiology.
2000;86:280-4.
2. Acevedo M, Pearce GL, Kottke-Marchant K, Sprecher DL. Elevated fibrinogen
and homocysteine levels enhance the risk of mortality in patients from a high-
risk preventive cardiology clinic. Arterioscler Thromb Vasc Biol. 2002;22:1042-
5.
3. Agewall S, Wikstrand J, Suurkula M, Tengborn L, Fagerberg B. Carotid artery
wall morphology, haemostatic factors and cardiovascular disease. An ultrasound
study in men at high and low risk for atherosclerotic disease. Blood Coagulation
& Fibrinolysis. 1994;5:895-904.
4. Anderson GL, Limacher M, Assaf AR, Bassford T, Beresford SAA, Black H,
Bonds D, Brunner R, Brzyski R, Caan B, Chlebowski R, Curb D, Gass M, Hays
J, Heiss G, Hendrix S, Howard BV, Hsia J, Hubbell A, Jackson R, Johnson KC,
Judd H, Kotchen JM, Kuller L, LaCroix AZ, Lane D, Langer RD, Lasser N,
Lewis CE, Manson J, Margolis K, Ockene J, O'Sullivan MJ, Phillips L, Prentice
RL, Ritenbaugh C, Robbins J, Rossouw JE, Sarto G, Stefanick ML, Van Horn L,
Wactawski-Wende J, Wallace R, Wassertheil-Smoller S, Women's Health
Initiative Steering C. Effects of conjugated equine estrogen in postmenopausal
women with hysterectomy: the Women's Health Initiative randomized controlled
trial.[see comment]. JAMA. 2004;291:1701-12.
5. Angerer P, Stork S, Kothny W, Schmitt P, von Schacky C. Effect of oral
postmenopausal hormone replacement on progression of atherosclerosis : a
randomized, controlled trial. Arterioscler Thromb Vasc Biol. 2001;21:262-8.
6. Anonymous. Effects of estrogen or estrogen/progestin regimens on heart disease
risk factors in postmenopausal women. The Postmenopausal Estrogen/Progestin
Interventions (PEPI) Trial. The Writing Group for the PEPI Trial. JAMA.
1995;273:199-208.
7. Auer J, Rammer M, Berent R, Weber T, Lassnig E, Eber B. Body iron stores
and coronary atherosclerosis assessed by coronary angiography. Nutrition
Metabolism & Cardiovascular Diseases. 2002;12:285-90.
139
8. Bielak LF, Klee GG, Sheedy PF, 2nd, Turner ST, Schwartz RS, Peyser PA.
Association of fibrinogen with quantity of coronary artery calcification
measured by electron beam computed tomography.[erratum appears in
Arterioscler Thromb Vasc Biol 2001 Apr;21(4):636]. Arteriosclerosis,
Thrombosis & Vascular Biology. 2000;20:2167-71.
9. Blankenhorn DH, Azen SP, Kramsch DM, Mack WJ, Cashin-Hemphill L, Hodis
HN, DeBoer LW, Mahrer PR, Masteller MJ, Vailas LI, Alaupovic P, Hirsch LJ,
Group MR. Coronary angiographic changes with lovastatin therapy. The
Monitored Atherosclerosis Regression Study (MARS). Annals of Internal
Medicine. 1993;119:969-76.
10. Bonithon-Kopp C, Scarabin PY, Taquet A, Touboul PJ, Malmejac A, Guize L.
Risk factors for early carotid atherosclerosis in middle-aged French women.
Arteriosclerosis & Thrombosis. 1991;11:966-72.
11. Brett KM, y C. Hormone replacement therapy: Knowledge and use in the
United States. Hyattsville, MD: National Center for Health Statistics; 2001.
12. Brown NJ, Abbas A, Byrne D, Schoenhard JA, Vaughan DE. Comparative
effects of estrogen and angiotensin-converting enzyme inhibition on
plasminogen activator inhibitor-1 in healthy postmenopausal women.
Circulation. 2002;105:304-9.
13. Chandler WL. The human fibrinolytic system. Crit Rev Oncol Hematol.
1996;24:27-45.
14. Chandler WL, Alessi MC, Aillaud MF, Henderson P, Vague P, Juhan-Vague I.
Clearance of tissue plasminogen activator (TPA) and TPA/plasminogen
activator inhibitor type 1 (PAI-1) complex: relationship to elevated TPA antigen
in patients with high PAI-1 activity levels. Circulation. 1997;96:761-8.
15. Chandler WL, Jascur ML, Henderson PJ. Measurement of different forms of
tPA in plasma. Clin Chem. 2000;46:38-46.
16. Chapman CML, Beilby JP, McQuillan BM, Thompson PL, Hung J. Monocyte
count, but not C-reactive protein or interleukin-6, is an independent risk marker
for subclinical carotid atherosclerosis. Stroke. 2004;35:1619-24.
17. Clarkson TB, Anthony MS, Morgan TM. Inhibition of postmenopausal
atherosclerosis progression: a comparison of the effects of conjugated equine
estrogens and soy phytoestrogens.[see comment]. Journal of Clinical
Endocrinology & Metabolism. 2001;86:41-7.
140
18. Clauss A. Gerinnungsphysiologische Schnellmethode zu Bestimmung des
Fibrinogens. Acta Haematol. 1957;17:237-245.
19. Cushman M, Meilahn EN, Psaty BM, Kuller LH, Dobs AS, Tracy RP. Hormone
replacement therapy, inflammation, and hemostasis in elderly women.
Arterioscler Thromb Vasc Biol. 1999;19:893-9.
20. Danesh J, Collins R, Appleby P, Peto R. Association of fibrinogen, C-reactive
protein, albumin, or leukocyte count with coronary heart disease: meta-analyses
of prospective studies. JAMA. 1998;279:1477-82.
21. de Ronde W, van der Schouw YT, Muller M, Grobbee DE, Gooren LJG, Pols
HAP, de Jong FH. Associations of sex-hormone-binding globulin (SHBG) with
non-SHBG-bound levels of testosterone and estradiol in independently living
men. Journal of Clinical Endocrinology & Metabolism. 2005;90:157-62.
22. Djousse L, Rothman KJ, Cupples LA, Levy D, Ellison RC. Serum albumin and
risk of myocardial infarction and all-cause mortality in the Framingham
Offspring Study. Circulation. 2002;106:2919-24.
23. Eichner JE, Moore WE, McKee PA, Schechter E, Reynolds DW, Qi H, Comp
PC. Fibrinogen levels in women having coronary angiography. American
Journal of Cardiology. 1996;78:15-8.
24. Elms MJ, Bunce IH, Bundesen PG, Rylatt DB, Webber AJ, Masci PP, Whitaker
AN. Measurement of crosslinked fibrin degradation products, an immunoassay
using monoclonal antibodies. Thromb Haemost. 1983;50:591-594.
25. Fischbach FT. Manual of Laboratory & Diagnostic Tests. Philadelphia, PA:
Lippincott Williams & Wilkins; 2004.
26. Folsom AR. Hemostatic risk factors for atherothrombotic disease: an
epidemiologic view. Thromb Haemost. 2001;86:366-73.
27. Francis RB, Jr., Kawanishi D, Baruch T, Mahrer P, Rahimtoola S, Feinstein DI.
Impaired fibrinolysis in coronary artery disease. American Heart Journal.
1988;115:776-80.
28. Gebara OC, Mittleman MA, Sutherland P, Lipinska I, Matheney T, Xu P, Welty
FK, Wilson PW, Levy D, Muller JE, et al. Association between increased
estrogen status and increased fibrinolytic potential in the Framingham Offspring
Study. Circulation. 1995;91:1952-8.
141
29. Godsland IF. Effects of postmenopausal hormone replacement therapy on lipid,
lipoprotein, and apolipoprotein (a) concentrations: analysis of studies published
from 1974-2000. Fertil Steril. 2001;75:898-915.
30. Goebelsmann U, Bernstein GS, Gale JA. Serum Gonadotropin testosterone
estradiol and estrone levels prior to and following bilateral vasectomy.
Vasectomy: Immunologic and pathophysiologic effects in animals and man.
1979;165.
31. Goebelsmann U, Horton R, Mestman JH, Arce JJ, Nagata Y, Nakamura RM,
Thorneycroft IH, Mishell DR, Jr. Male pseudohermaphroditism due to testicular
17 -hydroxysteroid dehydrogenase deficiency. Journal of Clinical
Endocrinology & Metabolism. 1973;36:867-79.
32. Grodstein F, Stampfer MJ, Manson JE, Colditz GA, Willett WC, Rosner B,
Speizer FE, Hennekens CH. Postmenopausal estrogen and progestin use and the
risk of cardiovascular disease.[see comment][erratum appears in N Engl J Med
1996 Oct 31;335(18):1406]. N Engl J Med. 1996;335:453-61.
33. Grotta JC, Yatsu FM, Pettigrew LC, Rhoades H, Bratina P, Vital D, Alam R,
Earls R, Picone C. Prediction of carotid stenosis progression by lipid and
hematologic measurements.[see comment]. Neurology. 1989;39:1325-31.
34. Halliwell B. Albumin--an important extracellular antioxidant? Biochemical
Pharmacology. 1988;37:569-71.
35. Hamsten A, de Faire U, Walldius G, Dahlen G, Szamosi A, Landou C,
Blomback M, Wiman B. Plasminogen activator inhibitor in plasma: risk factor
for recurrent myocardial infarction. Lancet. 1987;2:3-9.
36. Hamsten A, Wiman B, de Faire U, Blomback M. Increased plasma levels of a
rapid inhibitor of tissue plasminogen activator in young survivors of myocardial
infarction. N Engl J Med. 1985;313:1557-63.
37. Handa K, Kono S, Saku K, Sasaki J, Kawano T, Sasaki Y, Hiroki T, Arakawa
K. Plasma fibrinogen levels as an independent indicator of severity of coronary
atherosclerosis. Atherosclerosis. 1989;77:209-13.
38. Hansson GK. The stabilized plaque: will the dream come true? European Heart
Journal Supplements. 2001;3:C69-C75.
39. Hermiller JB, Cusma JT, Spero LA, Fortin DF, Harding MB, Bashore TM.
Quantitative and qualitative coronary angiographic analysis: review of methods,
utility, and limitations. Catheterization & Cardiovascular Diagnosis.
1992;25:110-31.
142
40. Herrington DM, Espeland MA, Crouse JR, 3rd, Robertson J, Riley WA,
McBurnie MA, Burke GL. Estrogen replacement and brachial artery flow-
mediated vasodilation in older women.[see comment]. Arteriosclerosis,
Thrombosis & Vascular Biology. 2001;21:1955-61.
41. Herrington DM, Reboussin DM, Brosnihan KB, Sharp PC, Shumaker SA,
Snyder TE, Furberg CD, Kowalchuk GJ, Stuckey TD, Rogers WJ, Givens DH,
Waters D. Effects of estrogen replacement on the progression of coronary-artery
atherosclerosis.[see comment]. N Engl J Med. 2000;343:522-9.
42. Hersh AL, Stefanick ML, Stafford RS. National use of postmenopausal hormone
therapy: annual trends and response to recent evidence.[see comment]. JAMA.
2004;291:47-53.
43. Hodis HN, Mack WJ, Azen SP, Lobo RA, Shoupe D, Mahrer PR, Faxon DP,
Cashin-Hemphill L, Sanmarco ME, French WJ, Shook TL, Gaarder TD, Mehra
AO, Rabbani R, Sevanian A, Shil AB, Torres M, Vogelbach KH, Selzer RH,
Women's Estrogen-Progestin Lipid-Lowering Hormone Atherosclerosis
Regression Trial Research G. Hormone therapy and the progression of coronary-
artery atherosclerosis in postmenopausal women.[see comment]. N Engl J Med.
2003;349:535-45.
44. Hodis HN, Mack WJ, Lobo RA, Shoupe D, Sevanian A, Mahrer PR, Selzer RH,
Liu Cr CR, Liu Ch CH, Azen SP, Estrogen in the Prevention of Atherosclerosis
Trial Research G. Estrogen in the prevention of atherosclerosis. A randomized,
double-blind, placebo-controlled trial. Ann Intern Med. 2001;135:939-53.
45. Holm P, Andersen HL, Andersen MR, Erhardtsen E, Stender S. The direct
antiatherogenic effect of estrogen is present, absent, or reversed, depending on
the state of the arterial endothelium. A time course study in cholesterol-clamped
rabbits. Circulation. 1999;100:1727-33.
46. Humes HDD, Herbert L.; Gardner, Laurence B.; Griffin, John W.; Harris,
Edward D.; Hazzard, William R.; King, Talmadge E.; Loriaux, D. Lynn;
Nabel, Elizabeth G.; Todd, Robert F.; Traber, Peter G. Kelley's Textbook of
Internal Medicine. Philadelphia, PA: Lippincott Williams & Wilkins; 2000.
47. Juhan-Vague I, Pyke SD, Alessi MC, Jespersen J, Haverkate F, Thompson SG.
Fibrinolytic factors and the risk of myocardial infarction or sudden death in
patients with angina pectoris. ECAT Study Group. European Concerted Action
on Thrombosis and Disabilities.[see comment]. Circulation. 1996;94:2057-63.
48. Kannel WB, Wolf PA, Castelli WP, D'Agostino RB. Fibrinogen and risk of
cardiovascular disease. The Framingham Study. JAMA. 1987;258:1183-6.
143
49. Karim R, Mack WJ, Lobo RA, Hwang J, Liu C-r, Liu C-h, Sevanian A, Hodis
HN. Determinants of the effect of estrogen on the progression of subclinical
atherosclerosis: Estrogen in the Prevention of Atherosclerosis Trial. Menopause.
2005;12:366-73.
50. Keber D. Mechanism of tissue plasminogen resease during venous occlusion.
Fibrinolysis. 1988;2(Suppl 2):96-103.
51. Kofoed SC, Wittrup HH, Sillesen H, Nordestgaard BG. Fibrinogen predicts
ischaemic stroke and advanced atherosclerosis but not echolucent, rupture-prone
carotid plaques: the Copenhagen City Heart Study.[see comment]. European
Heart Journal. 2003;24:567-76.
52. Koh KK, Horne MK, 3rd, Csako G, Waclawiw MA, Cannon RO, 3rd. Relation
of fibrinolytic potentiation by estrogen to coagulation pathway activation in
postmenopausal women. American Journal of Cardiology. 1999;83:466-9.
53. Koh KK, Mincemoyer R, Bui MN, Csako G, Pucino F, Guetta V, Waclawiw M,
Cannon RO, 3rd. Effects of hormone-replacement therapy on fibrinolysis in
postmenopausal women. N Engl J Med. 1997;336:683-90.
54. Kragh-Hansen U. Structure and ligand binding properties of human serum
albumin. Danish Medical Bulletin. 1990;37:57-84.
55. Kullo IJ, McConnell JP, Bailey KR, Kardia SL, Bielak LF, Peyser PA, Sheedy
PF, 2nd, Boerwinkle E, Turner ST. Relation of C-reactive protein and fibrinogen
to coronary artery calcium in subjects with systemic hypertension. American
Journal of Cardiology. 2003;92:56-8.
56. Langer RD, Pradhan AD, Lewis CE, Manson JE, Rossouw JE, Hendrix SL,
LaCroix AZ, Ridker PM. Baseline associations between postmenopausal
hormone therapy and inflammatory, haemostatic, and lipid biomarkers of
coronary heart disease. The Women's Health Initiative Observational Study.
Thrombosis & Haemostasis. 2005;93:1108-16.
57. Lassila R, Peltonen S, Lepantalo M, Saarinen O, Kauhanen P, Manninen V.
Severity of peripheral atherosclerosis is associated with fibrinogen and
degradation of cross-linked fibrin. Arteriosclerosis & Thrombosis.
1993;13:1738-42.
58. Levenson J, Giral P, Razavian M, Gariepy J, Simon A. Fibrinogen and silent
atherosclerosis in subjects with cardiovascular risk factors. Arteriosclerosis,
Thrombosis & Vascular Biology. 1995;15:1263-8.
144
59. Lindoff C, Petersson F, Lecander I, Martinsson G, Astedt B. Passage of the
menopause is followed by haemostatic changes. Maturitas. 1993;17:17-22.
60. Lip GY, Blann AD. von Willebrand factor and its relevance to cardiovascular
disorders. British Heart Journal. 1995;74:580-3.
61. Losner S, Volk BW, Wilensky ND. Fibrinogen concentration in acute
myocardial infarction; comparison of the clot density determination of
fibrinogen with the erythrocyte sedimentation rate. A M A Archives of Internal
Medicine. 1954;93:231-45.
62. Losordo DW, Kearney M, Kim EA, Jekanowski J, Isner JM. Variable
expression of the estrogen receptor in normal and atherosclerotic coronary
arteries of premenopausal women. Circulation. 1994;89:1501-10.
63. Lowe GD, Drummond MM, Lorimer AR, Hutton I, Forbes CD, Prentice CR,
Barbenel JC. Relation between extent of coronary artery disease and blood
viscosity. British Medical Journal. 1980;280:673-4.
64. Lowe GDO. Hormone replacement therapy and cardiovascular disease:
increased risks of venous thromboembolism and stroke, and no protection from
coronary heart disease. Journal of Internal Medicine. 2004;256:361-74.
65. Mack WJ, LaBree L, Liu C, Selzer RH, Hodis HN. Correlations between
measures of atherosclerosis change using carotid ultrasonography and coronary
angiography. Atherosclerosis. 2000;150:371-9.
66. Madsen JS, Kristensen SR, Gram J, Bladbjerg EM, Henriksen FL, Christensen
K, Jespersen J. Positive impact of hormone replacement therapy on the
fibrinolytic system: a long-term randomized controlled study in healthy
postmenopausal women. J Thromb Haemost. 2003;1:1984-91.
67. Manson JE, Hsia J, Johnson KC, Rossouw JE, Assaf AR, Lasser NL, Trevisan
M, Black HR, Heckbert SR, Detrano R, Strickland OL, Wong ND, Crouse JR,
Stein E, Cushman M, Women's Health Initiative I. Estrogen plus progestin and
the risk of coronary heart disease.[see comment]. New England Journal of
Medicine. 2003;349:523-34.
68. Maresca G, Di Blasio A, Marchioli R, Di Minno G. Measuring plasma
fibrinogen to predict stroke and myocardial infarction: an update. Arterioscler
Thromb Vasc Biol. 1999;19:1368-77.
145
69. Margaglione M, Di Minno G, Grandone E, Vecchione G, Celentano E, Cappucci
G, Grilli M, Simone P, Panico S, Mancini M. Abnormally high circulation levels
of tissue plasminogen activator and plasminogen activator inhibitor-1 in patients
with a history of ischemic stroke. Arterioscler Thromb. 1994;14:1741-5.
70. Meade TW, Ruddock V, Stirling Y, Chakrabarti R, Miller GJ. Fibrinolytic
activity, clotting factors, and long-term incidence of ischaemic heart disease in
the Northwick Park Heart Study.[see comment]. Lancet. 1993;342:1076-9.
71. Mertens I, Van Gaal LF. Obesity, haemostasis and the fibrinolytic system. Obes
Rev. 2002;3:85-101.
72. Mukherjee M, De Lorenzo F, Kadziola Z, Rutlin A, Ranlall N, Sembhi K,
Dawson G, Kakkar VV. Correlation of circulating 17beta-oestradiol with
haemostatic factors in healthy postmenopausal women.[erratum appears in
Haemostasis 2000 May-Jun;30(3):138 Note: DeLorenzo, F [corrected to De
Lorenzo, F]]. Haemostasis. 1999;29:336-42.
73. Nabulsi AA, Folsom AR, White A, Patsch W, Heiss G, Wu KK, Szklo M.
Association of hormone-replacement therapy with various cardiovascular risk
factors in postmenopausal women. The Atherosclerosis Risk in Communities
Study Investigators. N Engl J Med. 1993;328:1069-75.
74. Nakahara K, Kazahaya Y, Shintani Y, Yamazumi K, Eguchi Y, Koga S, Wada
H, Matsuda M. Measurement of soluble fibrin monomer-fibrinogen complex in
plasmas derived from patients with various underlying clinical situations.
Thrombosis & Haemostasis. 2003;89:832-6.
75. Newby DE, McLeod AL, Uren NG, Flint L, Ludlam CA, Webb DJ, Fox KA,
Boon NA. Impaired coronary tissue plasminogen activator release is associated
with coronary atherosclerosis and cigarette smoking: direct link between
endothelial dysfunction and atherothrombosis. Circulation. 2001;103:1936-41.
76. Paramo JA, Orbe J, Fernandez J. Fibrinolysis/proteolysis balance in stable
angina pectoris in relation to angiographic findings. Thrombosis & Haemostasis.
2001;86:636-9.
77. Park JS, Jung HH, Yang WS, Kim SB, Min WK, Chi HS. Effects of hormonal
replacement therapy on lipid and haemostatic factors in post-menopausal ESRD
patients. Nephrol Dial Transplant. 2000;15:1835-40.
78. Pearson TA. Coronary arteriography in the study of the epidemiology of
coronary artery disease. Epidemiologic Reviews. 1984;6:140-66.
146
79. Post MS, van der Mooren MJ, van Baal WM, Blankenstein MA, Merkus
HMWM, Kroeks MVAM, Franke HR, Kenemans P, Stehouwer CDA. Effects of
low-dose oral and transdermal estrogen replacement therapy on hemostatic
factors in healthy postmenopausal women: a randomized placebo-controlled
study. American Journal of Obstetrics & Gynecology. 2003;189:1221-7.
80. Post WS, Goldschmidt-Clermont PJ, Wilhide CC, Heldman AW, Sussman MS,
Ouyang P, Milliken EE, Issa JP. Methylation of the estrogen receptor gene is
associated with aging and atherosclerosis in the cardiovascular system.
Cardiovascular Research. 1999;43:985-91.
81. Pradhan AD, LaCroix AZ, Langer RD, Trevisan M, Lewis CE, Hsia JA,
Oberman A, Kotchen JM, Ridker PM. Tissue plasminogen activator antigen and
D-dimer as markers for atherothrombotic risk among healthy postmenopausal
women. Circulation. 2004;110:292-300.
82. Qu W, Le TT, Azen SP, Xiang M, Wong ND, Doherty TM, Detrano RC. Value
of coronary artery calcium scanning by computed tomography for predicting
coronary heart disease in diabetic subjects. Diabetes Care. 2003;26:905-10.
83. Ridker PM, Brown NJ, Vaughan DE, Harrison DG, Mehta JL. Established and
emerging plasma biomarkers in the prediction of first atherothrombotic events.
Circulation. 2004;109:6-19.
84. Ridker PM, Hennekens CH, Stampfer MJ, Manson JE, Vaughan DE.
Prospective study of endogenous tissue plasminogen activator and risk of
stroke.[see comment]. Lancet. 1994;343:940-3.
85. Rosano GMC, Vitale C, Fini M. Hormone replacement therapy and
cardioprotection: what is good and what is bad for the cardiovascular system?
Annals of the New York Academy of Sciences. 2006;1092:341-8.
86. Rosenfeld ME, Kauser K, Martin-McNulty B, Polinsky P, Schwartz SM,
Rubanyi GM. Estrogen inhibits the initiation of fatty streaks throughout the
vasculature but does not inhibit intra-plaque hemorrhage and the progression of
established lesions in apolipoprotein E deficient mice. Atherosclerosis.
2002;164:251-9.
87. Rosner W. Plasma steroid-binding proteins. Endocrinology & Metabolism
Clinics of North America. 1991;20:697-720.
88. Rosner W, Hryb DJ, Khan MS, Nakhla AM, Romas NA. Sex hormone-binding
globulin mediates steroid hormone signal transduction at the plasma membrane.
Journal of Steroid Biochemistry & Molecular Biology. 1999;69:481-5.
147
89. Sakata T, Mannami T, Baba S, Kokubo Y, Kario K, Okamoto A, Kumeda K,
Ohkura N, Katayama Y, Miyata T, Tomoike H, Kato H. Potential of free-form
TFPI and PAI-1 to be useful markers of early atherosclerosis in a Japanese
general population (the Suita Study): association with the intimal-medial
thickness of carotid arteries. Atherosclerosis. 2004;176:355-60.
90. Salomaa V, Stinson V, Kark JD, Folsom AR, Davis CE, Wu KK. Association of
fibrinolytic parameters with early atherosclerosis. The ARIC Study.
Atherosclerosis Risk in Communities Study. Circulation. 1995;91:284-90.
91. Salonen R, Seppanen K, Rauramaa R, Salonen JT. Prevalence of carotid
atherosclerosis and serum cholesterol levels in eastern Finland. Arteriosclerosis.
1988;8:788-92.
92. Scarabin PY, Vissac AM, Kirzin JM, Bourgeat P, Amiral J, Agher R, Guize L.
Population correlates of coagulation factor VII. Importance of age, sex, and
menopausal status as determinants of activated factor VII. Arteriosclerosis,
Thrombosis & Vascular Biology. 1996;16:1170-6.
93. Schmitz-Huebner U, Thompson SG, Balleisen L, Fechtrup C, Grosse-Heitmeyer
W, Kirchhof B, Most E, Muller US, Seiffert C, Seiffert D, et al. Lack of
association between haemostatic variables and the presence or the extent of
coronary atherosclerosis. British Heart Journal. 1988;59:287-91.
94. Schneidau A, Harrison MJ, Hurst C, Wilkes HC, Meade TW. Arterial disease
risk factors and angiographic evidence of atheroma of the carotid artery.[see
comment]. Stroke. 1989;20:1466-71.
95. Slater CC, Hodis HN, Mack WJ, Shoupe D, Paulson RJ, Stanczyk FZ. Markedly
elevated levels of estrone sulfate after long-term oral, but not transdermal,
administration of estradiol in postmenopausal women. Menopause. 2001;8:200-
3.
96. Slater CC, Zhang C, Hodis HN, Mack WJ, Boostanfar R, Shoupe D, Paulson RJ,
Stanczyk FZ. Comparison of estrogen and androgen levels after oral estrogen
replacement therapy. Journal of Reproductive Medicine. 2001;46:1052-6.
97. Small M, Lowe GD, Beattie JM, Hutton I, Lorimer AR, Forbes CD. Severity of
coronary artery disease and basal fibrinolysis. Haemostasis. 1987;17:305-11.
98. Sodergard R, Backstrom T, Shanbhag V, Carstensen H. Calculation of free and
bound fractions of testosterone and estradiol-17 beta to human plasma proteins
at body temperature. Journal of Steroid Biochemistry. 1982;16:801-10.
148
99. Sowers M, Derby C, Jannausch ML, Torrens JI, Pasternak R. Insulin resistance,
hemostatic factors, and hormone interactions in pre- and perimenopausal
women: SWAN. Journal of Clinical Endocrinology & Metabolism.
2003;88:4904-10.
100. Steiner AZ, Hodis HN, Lobo RA, Shoupe D, Xiang M, Mack WJ.
Postmenopausal oral estrogen therapy and blood pressure in normotensive and
hypertensive subjects: the Estrogen in the Prevention of Atherosclerosis Trial.
Menopause. 2005;12:728-33.
101. Tejada C, Strong JP, Montenegro MR, Restrepo C, Solberg LA. Distribution of
coronary and aortic atherosclerosis by geographic location, race, and sex.
Laboratory Investigation. 1968;18:509-26.
102. Thogersen AM, Jansson JH, Boman K, Nilsson TK, Weinehall L, Huhtasaari F,
Hallmans G. High plasminogen activator inhibitor and tissue plasminogen
activator levels in plasma precede a first acute myocardial infarction in both men
and women: evidence for the fibrinolytic system as an independent primary risk
factor. Circulation. 1998;98:2241-7.
103. Topol EJ, Califf, Robert M., Isner, Jeffrey M., Prystowsky, Eric N., Swain,
Judith L., Thomas, James D., Thompson, Paul D., Young, James B. Textbook of
Cardiovascular Medicine. Philadelphia, PA: Lippincott Williams & Wilkins;
2002.
104. van Baal WM, Kenemans P, Emeis JJ, Schalkwijk CG, Mijatovic V, van der
Mooren MJ, Vischer UM, Stehouwer CD. Long-term effects of combined
hormone replacement therapy on markers of endothelial function and
inflammatory activity in healthy postmenopausal women. Fertility & Sterility.
1999;71:663-70.
105. Vapaatalo H, Mervaala E. Clinically important factors influencing endothelial
function. Medical Science Monitor. 2001;7:1075-85.
106. Vermeulen A, Verdonck L, Kaufman JM. A critical evaluation of simple
methods for the estimation of free testosterone in serum. Journal of Clinical
Endocrinology & Metabolism. 1999;84:3666-72.
107. Vigen C, Hodis H, Chandler W, Lobo R, Mack W. Postmenopausal oral
estrogen therapy affects hemostatic factors, but does not account for reduction in
the progression of subclinical atherosclerosis. Journal of Thrombosis and
Haemostasis. 2007;5:1201-1208.
149
108. Vigen C, Hodis H, Chandler W, Stanczyk F, Lobo R, Mack W. Levels of
circulating hemostatic factors are related to postmenopausal estrogen therapy,
but not to circulating estrogen levels. Unpublished work.
109. Wenger NK. Cardioprotection for the postmenopausal women: HERS results
and their implications. Preventive Cardiology. 1998;1:9-11.
110. Wexler L, Brundage B, Crouse J, Detrano R, Fuster V, Maddahi J, Rumberger J,
Stanford W, White R, Taubert K. Coronary artery calcification:
pathophysiology, epidemiology, imaging methods, and clinical implications. A
statement for health professionals from the American Heart Association. Writing
Group. Circulation. 1996;94:1175-92.
111. Wilcox JG, Hwang J, Hodis HN, Sevanian A, Stanczyk FZ, Lobo RA.
Cardioprotective effects of individual conjugated equine estrogens through their
possible modulation of insulin resistance and oxidation of low-density
lipoprotein. Fertility & Sterility. 1997;67:57-62.
112. Wilhelmsen L, Svardsudd K, Korsan-Bengtsen K, Larsson B, Welin L, Tibblin
G. Fibrinogen as a risk factor for stroke and myocardial infarction. N Engl J
Med. 1984;311:501-5.
113. Willeit J, Kiechl S. Prevalence and risk factors of asymptomatic extracranial
carotid artery atherosclerosis. A population-based study. Arteriosclerosis &
Thrombosis. 1993;13:661-8.
114. Williams JK, Anthony MS, Honore EK, Herrington DM, Morgan TM, Register
TC, Clarkson TB. Regression of atherosclerosis in female monkeys.
Arteriosclerosis, Thrombosis & Vascular Biology. 1995;15:827-36.
115. Wu S, Ruan Y, Zhu X, Lai W. Estrogen receptors and the activity of nitric oxide
synthase in the artery of female rats receiving hormone replacement therapy.
Hormone Research. 2000;53:144-7.
116. Yildiz BO, Haznedarolu IC, Kirazli S, Bayraktar M. Global fibrinolytic capacity
is decreased in polycystic ovary syndrome, suggesting a prothrombotic state.
Journal of Clinical Endocrinology & Metabolism. 2002;87:3871-5.
117. Yock PG, Linker DT. Intravascular ultrasound. Looking below the surface of
vascular disease.[comment]. Circulation. 1990;81:1715-8.
150
118. Zheng ZJ, Sharrett AR, Chambless LE, Rosamond WD, Nieto FJ, Sheps DS,
Dobs A, Evans GW, Heiss G. Associations of ankle-brachial index with clinical
coronary heart disease, stroke and preclinical carotid and popliteal
atherosclerosis: the Atherosclerosis Risk in Communities (ARIC) Study.
Atherosclerosis. 1997;131:115-25.
119. Zoellner H, Hofler M, Beckmann R, Hufnagl P, Vanyek E, Bielek E, Wojta J,
Fabry A, Lockie S, Binder BR. Serum albumin is a specific inhibitor of
apoptosis in human endothelial cells. Journal of Cell Science. 1996;109:2571-
80.

Abstract (if available)

Abstract Numerous cardiovascular risk factors have been identified, and lifestyle and drug interventions are available for modification of many of these risk factors. Nevertheless, cardiovascular disease remains the leading cause of death among both men and women in the industrialized world. This has spurred interest in identifying novel risk factors and markers in order to reduce cardiovascular disease morbidity and mortality.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Hormone therapy timing hypothesis and atherosclerosis

PDF

Sex hormones and atherosclerosis in postmenopausal women

PDF

Subclinical atherosclerosis in overweight Latino youth: influence of cardiometabolic risk factors

PDF

Cardiovascular disease risk factors and cognitive function

PDF

Study of risk factors of B-cell non-Hodgkin lymphoma (NHL) in the California Teachers Study

PDF

Non-steroidal anti-inflammatory drugs, anti-acid drugs and passive smoking as impact factors for esophageal and gastric adenocarcinomas

PDF

Association between cardiovascular risk factors and coronary CT measures of coronary atherosclerosis in healthy postmenopausal women

PDF

Dietary fats, atherosclerosis, and cardiovascular disease

PDF

Intrapersonal and environmental factors associated with Chinese youth alcohol use experimentation and binge drinking behaviors

PDF

The metabolic syndrome in overweight Latino youth: influence of dietary intake and associated risk for Type 2 diabetes

PDF

Association between endogenous sex hormone levels and cognition: the Women’s Isoflavone Soy Health (WISH) trial

PDF

The effects of tobacco exposure on hormone levels and breast cancer risk among young women

PDF

Motivation and the meanings of health behavior as factors associated with eating behavior in Latino youth

PDF

Factors and correlates of sexual behaviors among young adults from continuation high schools

PDF

Psychosocial and cultural factors in the primary prevention of melanoma targeted to multiethnic children

PDF

Competing risks: the role of the perceived consequences of refusing to share injection equipment among injection drug users

PDF

Sex differences in iron overload

PDF

Objectively measured physical activity and related factors in minority youth

PDF

Physical activity and sex hormone levels in postmenopausal women

PDF

Psychosocial and behavioral ractors associated with emotional eating in adolescents

Asset Metadata

Creator Vigen, Cheryl (author)

Core Title Associations between sex steroid hormones, hemostatic factors and atherosclerosis

School Keck School of Medicine

Degree Doctor of Philosophy

Degree Program Preventive Medicine (Health Behavior)

Publication Date 07/13/2007

Defense Date 04/25/2007

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag atherosclerosis,hemostatic factors,Hormones,OAI-PMH Harvest

Language English

Advisor Mack, Wendy J. (committee chair), Hodis, Howard Neil (committee member), Kuenzli, Nino (committee member), Levine, Alexandra (committee member), Meiselman, Herbert J. (committee member)

Creator Email vigen@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-m617

Unique identifier UC1437332

Identifier etd-Vigen-20070713 (filename),usctheses-m40 (legacy collection record id),usctheses-c127-517378 (legacy record id),usctheses-m617 (legacy record id)

Legacy Identifier etd-Vigen-20070713.pdf

Dmrecord 517378

Document Type Dissertation

Rights Vigen, Cheryl

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Repository Name Libraries, University of Southern California

Repository Location Los Angeles, California

Repository Email cisadmin@lib.usc.edu