Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
Effect of soy isoflavones on anthropometric and metabolic measurements in postmenopausal women
(USC Thesis Other)
Effect of soy isoflavones on anthropometric and metabolic measurements in postmenopausal women
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
EFFECT OF SOY ISOFLAVONES ON ANTHROPOMETRIC AND
METABOLIC MEASUREMENTS IN POSTMENOPAUSAL WOMEN
by
Jun Wang
A Thesis Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
MASTER OF SCIENCE
(APPLIED BIOSTATISTICS AND EPIDEMIOLOGY)
May 2011
Copyright 2011 Jun Wang
ii
ACKNOWLEDGMENTS
I would like to express my sincere gratitude to Dr. Wendy Mack for her invaluable
inspiration and guidance throughout my thesis experience. I am thankful to Dr. Stanley
Azen and Dr. Howard Hodis for their support in preparing the thesis. I am also thankful
to Naoko Kono for her help in preparing and analyzing the data. I would like to thank my
parents and husband for their encouragement and support.
iii
TABLE OF CONTENTS
Acknowledgments ii
Table of contents iii
List of tables iv
List of figures vi
Abbreviations vii
Abstract viii
Chapter 1: Background and introduction 1
1.1 Estrogen and Ht 1
1.2 Isoflavones 2
1.3 Animal studies 4
1.4 Observational studies in humans 5
1.5 Interventional studies in humans 5
1.6 Equol 7
1.7 Objectives 8
Chapter 2: Materials and methods 10
2.1 Study design 10
2.2 Randomization 11
2.3 Intervention 12
2.4 Baseline and follow-up 12
2.5 Laboratory assays 13
2.6 Statistical analysis 14
Chapter 3: Results 15
3.1 Subject characteristics 15
3.2 Anthropometric measurements 18
3.4 Equol 25
3.5 Plasma IFLS 27
Chapter 4: Discussion 33
Bibliography 37
Appendix 42
iv
LIST OF TABLES
Table 1. WISH demographic and baseline characteristics by treatment 16
Table 2. Baseline anthropometric and metabolic characteristics stratified by equol
producer subgroups 17
Table 3. Mean change in BMI (kg/m
2
) by treatment 18
Table 4. Mean change in weight (lbs) by treatment 19
Table 5. Mean change in hip circumference (inches) by treatment 19
Table 6. Mean change in waist circumference (inches) by treatment 20
Table 7. Mean change in waist to hip ratio by treatment 20
Table 8. Mean change in total cholesterol (mg/dL) by treatment 21
Table 9. Mean change in HDL-cholesterol (mg/dL) by treatment 22
Table 10. Mean change in LDL-cholesterol (mg/dL) by treatment 23
Table 11. Mean change in triglycerides (mg/dL) by treatment 24
Table 12. Mean change in glucose (mg/dL) by treatment 25
Table 13. Mean change in anthropometric and metabolic variables by equol producer
subgroup 26
Table 14. association among plasma IFLs and change in HDL-cholesterol 28
Table 15. Association among plasma IFLs and change in total cholesterol 28
Table 16. Association among plasma IFLs and change in BMI 29
Table 17. Association among plasma IFLs and change in LDL-cholesterol 29
Table 18. Association among plasma IFLs and change in triglycerides 30
Table 19. Association among plasma IFLs and change in glucose 30
v
Table 20. Association among plasma IFLs and change in weight 31
Table 21. Association among plasma IFLs and change in waist to hip ratio 31
Table 22. Association among plasma IFLs and change in hip circumference 32
Table 23. Association among plasma IFLs and change in waist circumference 32
Table 24. Association between change in plasma IFLs and change in HDL-cholesterol 42
Table 25. Association between change in plasma IFLs and change in total cholesterol 42
Table 26. Association between change in plasma IFLs and change in BMI 43
Table 27. Association between change in plasma IFLs and change in LDL-cholesterol 43
Table 28. Association between change in plasma IFLs and change in triglycerides 44
Table 29. Association between change in plasma IFLs and change in glucose 44
Table 30. Association between change in plasma IFLs and change in weight 45
Table 31. Association between change in plasma IFLs and change in waist to hip ratio 45
Table 32. Association between change in plasma IFLs and change in hip circumference46
Table 33. Association between change in plasma IFLs and change in waist circumference
46
Table 34. Previous interventional studies on the effect of isoflavones in humans 47
vi
LIST OF FIGURES
Figure 1. Chemical structures of S-equol and R-equol (Setchell, Clerici et al. 2005) 3
Figure 2. Flow chart of WISH 11
vii
ABBREVIATIONS
BMI: Body Mass Index
CHOL: Total Cholesterol
CIMT (Carotid Intima-Media Thickness)
ER: Estrogen Receptor
HDL-cholesterol: High-density Lipoprotein Cholesterol
HT: Hormone Therapy
IFL: Isoflavone level
ISP: Soy Isoflavone Protein
LDL: Low-density Lipoprotein Cholesterol
TG: Triglycerides
WISH: Women’s Isoflavone Soy Health
viii
ABSTRACT
In this single-center, randomized, double-blinded, placebo-controlled clinical trial,
350 postmenopausal women were randomized to either soy isoflavone or placebo groups
and followed for 2.5 years. The soy isoflavone group received 25g soy protein containing
85mg soy isoflavones daily, and the placebo group received matched placebo.
Generalized estimating equations with identity link function and exchangeable
correlation structure were conducted to examine the treatment effects on anthropometric
and metabolic parameters in total subjects or among subgroups defined by equol producer,
time since menopause, age and ethnicity. The association between plasma isoflavone
levels and changes in anthropometric and metabolic parameters were also examined.
The soy isoflavone group did not differ from the placebo group in mean change of
LDL-cholesterol, glucose, total cholesterol, or anthropometric measurements. Compared
to placebo, the mean increase of HDL-cholesterol was marginally higher in the soy
isoflavones group (p=0.05), the difference was significant in white women and those who
experienced menopause less than 5 years ago (both p=0.04). The mean increase of
triglycerides was marginally lower (p=0.05) in the ISP group compared to placebo
(p=0.05), the difference was significant in white women (p=0.03) and women aged 56 to
60 years old (p=0.01). The mean change of triglycerides in the non-equol producer and
the consistent equol producer subgroups were significantly lower than the placebo group
(both p=0.04). Plasma dihydrogenistein was negatively associated with hip circumference
(β=-0.56; p=0.01) but positively associated with waist circumference (β=0.53; p=0.02).
ix
Plasma genistein level was also negatively associated with hip circumference (β=-0.16;
p=0.01).
This analysis indicates that daily consumption of 25g soy protein containing 85mg
isoflavones has no overall effect on anthropometric measurements but has potentially
beneficial effects on lipids in postmenopausal women. However, plasma isoflavone levels
are associated with hip and waist circumference. Equol-producing capability modifies the
effect of isoflavones on triglycerides.
1
CHAPTER 1: BACKGROUND AND INTRODUCTION
1.1 ESTROGEN AND HT
Estrogens are a group of steroid compounds that function as the primary female sex
hormone. Estrogen levels are higher in women of reproductive age and generally
decrease between the age of 45 and 55. In surgically postmenopausal women (whose
ovaries are removed), estrogen levels are also extremely low. In addition to the
development and expression of female secondary sex characteristics and regulation of the
menstrual cycle, estrogens affect the body in many other aspects, including body weight
and fat distribution (Genazzani and Gambacciani 2006). Modification of estrogen levels
is linked to central and peripheral expression of the leptin receptor gene (OR-Rb), which
mediates the weight-reducing effect of leptin (Meli, Pacilio et al. 2004). Decreased
estrogen levels may also lead to deleterious alterations in glucose tolerance and plasma
lipid levels (Sudhaa Sharma 2008). These changes can be decreased or even reversed by
hormone therapy (HT).
HT is commonly used to prevent menopausal symptoms (such as hot flashes,
irregular menstruation, and night sweats) that are caused by the decrease of endogenous
sex hormones at menopause. Estrogens, progesterone, progestins, and sometimes
testosterone are the main types of hormones comprising HT. Estrogens in the form of HT
have been shown to decrease the change of body fat distribution (Genazzani and
Gambacciani 2006). Exogenous estrogen together with progesterone have been shown to
decrease plasma concentrations of LDL-cholesterol and increase HDL-cholesterol
2
(Erberich, Alcantara et al. 2002). Despite those effects, researchers have focused on the
adverse effects associated with HT. Use of HT may increase the risk of cancers, including
breast and ovarian cancer (Association 2002; Beral, Bull et al. 2007), and can result in
endometrial hyperplasia, which is a precursor to endometrial cancer (Shields, Weiss et al.
1999). Because of these undesirable side effects, resistance to the use of estrogen therapy
is growing. The problems and uncertainties associated with traditional HT call for the use
of possibly safer natural products, in particular isoflavone-rich soy protein as a
postmenopausal therapeutic alternative.
1.2 ISOFLAVONES
Isoflavones are one of the main classes of phytoestrogens. Phytoestrogens are found
in various plants, including nuts, oil seeds, fruits, and vegetables. The highest
concentrations of isoflavones are found in soybeans and soy bean products. Because of
the structural similarity to estradiol (17β-estradiol), isoflavones can bind to estrogen
receptors (both ER-α and ER-β) to modulate gene transcription and cell-signaling
pathways (Hall, Rimbach et al. 2005). However the estrogenic activity of isoflavones is
weaker than estradiol (their proliferative activity is 10
-4
-10
-5
times compared to 17β-
estradiol) (Fritsche S 1999). Isoflavones not only bind with ERs but also with
peroxisome proliferator-activated receptor (PPAR) families, which are the primary
adipogenic transcription factors. An in vitro study demonstrated that daidzein decreases
adipogenesis at concentrations lower than 20μM and stimulates adipogenesis at
concentrations higher than 30μM (Dang and Lowik 2004).
The metabolism of isoflavones is a complex process of the digestive system and gut
microflora. After ingestion, isoflavones are hydrolyzed by intestinal glucosidases to
several components, including aglycones genistein, daidzein and glycitein
et al. 2003). These compounds
Although 17β-estradiol has equal affinities for
has a higher affinity for ER
intestinal microflora influence serum isoflavone profiles among individuals who take the
same dose of isoflavones. Intestinal bacteria can convert daidzein to equol, which
a nonsteroidal estrogen. Only about 30
can convert daidzein to equol
affinity for both ER-α and
3 position of the furan ring, thus there are 2 distinct enantiomeric forms: S
equol (Fig 1). Possibly because equol (S
gene expression may differ in equol producers and non
et al. 2005; Niculescu, Pop et al. 2007)
Figure 1. Chemical structures of S
The metabolism of isoflavones is a complex process of the digestive system and gut
microflora. After ingestion, isoflavones are hydrolyzed by intestinal glucosidases to
several components, including aglycones genistein, daidzein and glycitein
. These compounds bind to the estrogen receptor and initiate gene expression
has equal affinities for ER-β and ER-α, the isoflavone g
has a higher affinity for ER-β than ER-α (Kuiper, Enmark et al. 1996)
intestinal microflora influence serum isoflavone profiles among individuals who take the
same dose of isoflavones. Intestinal bacteria can convert daidzein to equol, which
a nonsteroidal estrogen. Only about 30-50% of individuals have intestinal bacteria that
can convert daidzein to equol (Kenneth D. R 2006). Equol is stable over time and has
α and ER-β (Cassidy 2006). Equol has a chiral carbon atom at
3 position of the furan ring, thus there are 2 distinct enantiomeric forms: S
equol (Fig 1). Possibly because equol (S-equol) has a relatively high affinity for ER
gene expression may differ in equol producers and non-equol producers
et al. 2005; Niculescu, Pop et al. 2007).
tructures of S-equol and R-equol (Setchell, Clerici et al. 2005)
3
The metabolism of isoflavones is a complex process of the digestive system and gut
microflora. After ingestion, isoflavones are hydrolyzed by intestinal glucosidases to
several components, including aglycones genistein, daidzein and glycitein (Bolego, Poli
bind to the estrogen receptor and initiate gene expression.
, the isoflavone genistein
(Kuiper, Enmark et al. 1996). In humans,
intestinal microflora influence serum isoflavone profiles among individuals who take the
same dose of isoflavones. Intestinal bacteria can convert daidzein to equol, which is also
50% of individuals have intestinal bacteria that
. Equol is stable over time and has
arbon atom at the C-
3 position of the furan ring, thus there are 2 distinct enantiomeric forms: S-equol and R-
equol) has a relatively high affinity for ER-β,
equol producers (Setchell, Clerici
rici et al. 2005)
4
1.3 ANIMAL STUDIES
Rodents are commonly used as an animal model for studying the effects of soy
isoflavones. Similar to postmenopausal women, experiments on ovariectomized female
mice showed that removal of the ovaries causes increases in adipose tissue, while
estrogen replacement reverses the increase (Wade, Gray et al. 1985). A 21-day study on
ovariectomized rats demonstrated that compared to the placebo group, a dose of 1500mg
daily genistein decreased body weight, parametrial fat pad and inguinal fat pad weights,
and increased apoptosis in inguinal fat; a dose of 150mg genistein showed no effect (Kim,
Nelson-Dooley et al. 2006). This study was consistent with a earlier study, which
demonstrated that in ovariectomized mice, 500-1500ppm dietary genistein decreases fat
pad weights dose dependently by 37-57%, but 300ppm genistein showed no effect (Naaz,
Yellayi et al. 2003). Some studies also reported that genistein supplementation decreased
triglycerides, total cholesterol and LDL-cholesterol, and glucose levels in the serum and
liver of rats (Aoyama, Fukui et al. 2000; Yang, Lee et al. 2006; Na, Ezaki et al. 2008).
Isoflavones may decrease fat accumulation and serum cholesterol levels by increasing
fatty acid oxidation, LDL receptor activity and mitochondrial enzyme activities (Kirk,
Sutherland et al. 1998; Mullen, Brown et al. 2004; Owen, Roach et al. 2004). However,
whether isoflavones alone or with other soy proteins contribute to the decrease of liver
and plasma lipids is still under discussion (Wagner, Schwenke et al. 2003). In applying
these findings to humans, the interspecies differences in isoflavone metabolism and
overall metabolic phenotype between rodents and humans cannot be neglected (Gu,
House et al. 2006).
5
1.4 OBSERVATIONAL STUDIES IN HUMANS
Isoflavones have been linked to many protective health benefits in women, including
amelioration of menopausal symptoms, cardiovascular disease, osteoporosis and breast
cancer (Ingram, Sanders et al. 1997; de Lemos 2001), but their effect on fat and lipids in
humans is still under discussion. Because of their structural similarity to estradiol, it is
possible that high concentrations of isoflavones may play a role in fat distribution and
lipids in humans.
One study examined the effect of isoflavones on anthropometric measurements in
postmenopausal women. It demonstrated that compared to women who did not consume
genistein daily, women with more than 1mg daily genistein intake had lower weight, total
fat mass, BMI and waist circumference. After adjustment for BMI, ethnicity, total energy
intake/day and total fiber intake/day, women with 1mg daily isoflavones had higher
HDL-cholesterol; however no associations were found with waist to hip ratio,
triglycerides, total cholesterol, LDL-cholesterol, or glucose (Goodman-Gruen and Kritz-
Silverstein 2001; Goodman-Gruen and Kritz-Silverstein 2003). This study was limited by
the assessment of the intake of dietary isoflavones which were assessed by self report.
1.5 INTERVENTIONAL STUDIES IN HUMANS (Summarized in Table 34)
Most of the previous trials have focused on lipids; however, the results have not
been consistent. In a 6-month trial among 66 hypercholesterolemic postmenopausal
women, subjects were randomized into control, ISP56 (55.6mg isoflavones daily) and
ISP90 (90mg isoflavones daily) groups. Compared to the control group, non-HDL
cholesterol decreased, and HDL-cholesterol and LDL receptor mRNA increased in both
6
ISP56 and ISP90 groups, while total cholesterol was not changed (Potter, Baum et al.
1998). In a randomized cross-over trial among 51 perimenopausal women, subjects
received placebo, 20g soy protein at one dose or split by two daily doses for each
treatment period of 6 weeks. Total cholesterol and LDL cholesterol were reduced in both
treatment groups with a placebo comparision, but no effects were found on HDL-
cholesterol or triglycerides (Washburn, Burke et al. 1999). In another 4-week randomized
trial among 23 Japanese perimenopausal women demonstrated that compared to the
placebo group, 61.8mg daily soy isoflavone extract significantly decreased total serum
cholesterol and LDL-cholesterol; no changes were found in HDL-cholesterol, VLDL-
cholesterol or triglycerides (Uesugi, Fukui et al. 2002). In a cross-over trial among 18
normocholesterolemic and mildly hypercholesterolemic postmenopausal women, 132mg
isoflavones daily reduced plasma LDL-cholesterol by 6.5% compared to placebo, but
65mg isoflavones daily did not significantly reduce the LDL-cholesterol concentration;
no effect was found on total cholesterol, HDL-cholesterol or triglycerides (Wangen,
Duncan et al. 2001). In a 12-week randomized trial among 216 postmenopausal women,
subjects were given 20g soy protein daily containing 160mg isoflavones or placebo. Total
cholesterol, LDL-cholesterol and LDL particle number in the isoflavone group were
significantly lower than the placebo group at 6 weeks, but the difference became
insignificant at 12 weeks (Allen, Becker et al. 2007). In a recent 12-week randomized
trial among 75 postmenopausal women, compared to the placebo group, 30gm soy
protein daily containing 60mg of isoflavones increased HDL-cholesterol levels and
reduced serum cholesterol, triglycerides, LDL-cholesterol levels, while 60mg soy
isoflavones extract daily only reduced the serum triglycerides level (H K Jassi 2010).
7
Some other studies found contradictory results on the effect of soy isoflavones in
postmenopausal women. In a cross-over trial among 64 postmenopausal women who
were treated for breast cancer, subjects were randomly given 114mg isoflavones or
placebo daily for 3 months, then after a 2 months washout period, subjects switched to
the alternate treatment for another 3 months. No effect was found on lipids or
apolipoproteins, although LDL-cholesterol increased during the isoflavone regimen and
decreased during the placebo regimen (Nikander, Tiitinen et al. 2004). Another trial
among 117 postmenopausal women found no effect of 50mg isoflavones (genistein to
daidzein ratio of 2:1) daily on plasma concentrations of lipids, glucose, or insulin (Hall,
Vafeiadou et al. 2006). A clinical trial among 202 postmenopausal women older than 60
years found no effect of 99mg isoflavones daily on plasma lipids in postmenopausal
women (Kreijkamp-Kaspers, Kok et al. 2004). In a recent trial among 100 obese
postmenopausal women, 80mg soy isoflavone extract daily showed no effect on lipids or
anthropometric measures (Llaneza, Gonzalez et al. 2010).
1.6 EQUOL
Since equol-producing ability varies between individuals, some studies examine the
influence of equol-producing capability on outcomes. A one-year trial among 54
Japanese postmenopausal women suggested that body fat accumulation depends on
equol-producing capacity. In this study, equol producers were defined as women who had
at least a 10% conversion rate of equol from daidzein after 96 hours incubation (Wu, Oka
et al. 2007). In a randomized crossover trial, 26 hypercholesterolaemic and hypertensive
volunteers were randomized to take 1000ml of yogurt containing 30g of soy protein or
8
1000ml of dairy milk for 5 weeks, then switched to the alternate treatment for another 5
weeks without a washout period. Total cholesterol, LDL-cholesterol, LDL/HDL ratio and
plasma triglycerides concentrations significantly decreased in equol producers (equol was
detected in plasma or urine) (Meyer, Larkin et al. 2004). However, in another trial, equol
producers were defined as subjects who had a log
10
value of urinary S-equol to daidzein
ratio above -1.75. LDL-cholesterol levels did not differ in equol producers compared to
non-equol producers (Alicia A Thorp 2008). In all these previous studies, there was no
unique standard in defining equol-producing status.
1.7 OBJECTIVES
To date, though there is some evidence that soy isoflavones have positive benefits
on lipids, the relationship between isoflavones and anthropometric and metabolic
measurements in postmenopausal women has not been clearly investigated. Most studies
have been conducted with a very small number of participants with a short intervention
period. Furthermore, no previous studies had examined the effect of plasma isoflavone
levels on anthropometric and metabolic measurements. Therefore, with a longer
intervention period and more subjects, we studied the effect of isoflavone-rich soy
protein supplementation on anthropometric and metabolic parameters in postmenopausal
women. The main objectives are:
1). To determine the effect of randomized treatment on anthropometric and metabolic
parameters a). with soy isoflavones versus placebo; b). in subgroups defined by equol
producer, time since menopause, age and ethnicity.
9
2). To determine the association between isoflavones levels measured in serum and
changes in anthropometric and metabolic parameters.
10
CHAPTER 2: MATERIALS AND METHODS
2.1 STUDY DESIGN
This analysis utilized data from the Women’s Isoflavone Soy Health (WISH) trial
conducted at the University of Southern California. The WISH was a single-center,
randomized, double-blinded, placebo-controlled clinical trial conducted from April, 2004
to March 2009. The aim of WISH was to determine the effect of isoflavone rich soy
protein supplementation on progression of atherosclerosis in postmenopausal women.
The data in this analysis were extracted from the WISH trial database.
Participants were postmenopausal women who had no clinical evidence of
preexisting cardiovascular disease. Postmenopausal status was defined as no vaginal
bleeding for at least one year and serum estradiol level less than 20pg/ml. Women were
excluded for the following reasons: 1. Clinical signs, symptoms, or personal history of
cardiovascular disease. 2. Diabetes mellitus or fasting serum glucose >126 mg/dL. 3.
Uncontrolled hypertension (diastolic blood pressure >110 mmHg). 4. Thyroid disease
(untreated). 5. Renal insufficiency (serum creatinine >2.0 mg/dL). 6. Life threatening
illness with prognosis <5 years. 7. Current use of hormone therapy. From 2004 to 2007,
three hundred and fifty women met the inclusion and exclusion criteria and agreed to
participate (Fig. 2). Written informed consent was obtained from all participants. The
study protocol was approved by the Institutional Review Board of the University of
Southern California.
11
Figure 2. Flow chart of WISH
2.2 RANDOMIZATION
Eligible subjects were randomly assigned into ISP and placebo groups within strata
defined by common carotid artery intima-media thickness (<0.75mm, >0.75mm).
Blocked randomization was used, with the block size masked to study staff and
investigators. The randomization scheme utilized a computerized random number
generator.
12
2.3 INTERVENTION
Participants were randomized to receive daily soy products or placebo. The soy
products were made with isolated soy protein consisting of 25g soy protein in total,
which contained 85mg aglycone weight naturally occurring isoflavones (150mg total
isoflavones) of genistein 45mg aglycone weight (80mg total weight), daidzein 35mg
aglycone weight (60mg total weight) and glycitein 5mg aglycone weight (10mg total
weight). The placebo products were milk protein isolate that contained no isoflavones.
Both soy and placebo products contained similar levels of macronutrients, minerals and
vitamins.
2.4 BASELINE AND FOLLOW-UP
Each subject attended two screening clinic visits and one baseline (randomization)
clinic visit separately to determine eligibility, complete randomization and collect
baseline outcome data. During the screening and baseline visits, demographic
characteristics were collected, including age, height, ethnicity, years since menopause
and type of menopause. Key measurements including weight, waist circumference, hip
circumference, total cholesterol, HDL-cholesterol, LDL-cholesterol, triglycerides, plasma
isoflavone concentrations (total isoflavones, daidzein, dihydrodaidzein, dihydrogenistein,
O-desmethylangolensin, equol, geinistein and glycitein) were also collected. Fasting
glucose level was assessed at the 2
nd
screening visit.
Follow-up visits occurred every month for the first 6 months and then every other
month thereafter for the remainder of the study. The initial 2.5-year treatment period was
extended to 3 years by the External Data and Safety Monitoring Board to increase the
13
chance of detecting treatment group differences on the primary endpoint. Interim
analyses of the primary trial endpoint were not performed. During follow up, data
regarding body weight, dietary intake, product compliance, non-study and nutritional
products, clinical adverse events and vital signs were ascertained at every clinic visit.
Glucose was measured every 12 months. The remaining measurements including waist
circumference, hip circumference, lipids (total cholesterol, total triglyceride, LDL-
cholesterol, VLDL-cholesterol, HDL-cholesterol, LDL-triglyceride, VLDL-triglyceride,
HDL-triglyceride) and plasma isoflavones concentrations were measured every 6 months.
2.5 LABORATORY ASSAYS
Fasting plasma lipids were measured using an enzymatic method under the CDC
Standardization Program. Plasma IFL levels were measured by HPLC with isotope
dilution electrospray ionization (negative mode) tandem mass spectrometry. Participants
fasted 8 hours before sample collections, and all specimens were immediately processed
and stored at -80°C. Plasma equol levels at post-randomization visits were used to
determine equol-producer status. Among the ISP participants, subjects were defined as
non equol producers (equol never>20nmol/L across post-randomization visits),
intermittent equol producers (equol>20nmol/L at some visits) and consistent equol
producers (equol>20nmol/L at all visits).
LDL was calculated using the Friedewald equation: LDL cholesterol = total
cholesterol - HDL cholesterol - (triglycerides/5). Waist to hip ratio = weight
circumference/ hip circumference. BMI = weight(kg) / (height(m))
2
14
2.6 STATISTICAL ANALYSIS
Data analysis was based on intent-to-treat, which was defined as all subjects that
were randomized and underwent baseline and at least one follow-up visit, regardless of
their compliance to their randomized treatments. Data were checked for normality before
further analysis. Since the distributions of variables did not depart from normal
distribution, no transformations were needed. To compare the treatment groups on
demographic and baseline characteristics, independent t-tests were conducted for
continuous variables, and Pearson chi-square tests were conducted for categorical
variables. One-way ANOVA was conducted to compare the baseline characteristics of
anthropometric and metabolic measurements among subgroups. To compare the mean
change in on-trial variables between treatment groups and among subgroups (defined by
equol producer, time since menopause, age and ethnicity), we utilized generalized
estimating equation, using an identity link function and exchangeable correlation
structure, and adjustment for randomization stratum. We also utilized two generalized
estimating equations models to assess the association between isoflavone levels and
change in metabolic outcome. One model included treatment group and randomization
stratum, and the second model only adjusted for randomization stratum. Both models
were performed with an identity link function and exchangeable correlation structure. All
statistical tests were conducted at a 2-sided alpha=0.05. Statistical analysis was
performed using SAS 9.1 software.
15
CHAPTER 3: RESULTS
3.1 SUBJECT CHARACTERISTICS
A total of 350 women met the inclusion criteria and were randomized; 161 (92%)
subjects in the placebo group and 161 (92%) subjects in the ISP group were followed for
at least one visit. 131 (74.9%) subjects in the placebo group and 141 (80.6%) subjects in
the ISP group completed the 2.5-year study (Fig. 1).
Participants ranged in age from 45 to 92 years at randomization; the mean(±SD) age
was 60.7(±6.7) years in the placebo group and 60.8(±7.0) years in the ISP group. Most
subjects were non-Hispanic white (63.7%), and most had experienced natural menopause
(89.7%). Demographic characteristics and baseline measurements are summarized in
Table 1. Treatment groups did not significantly differ in age, race, type of menopause, or
years since menopause. At baseline, treatment groups did not significantly differ in
anthropometric or metabolic measurements (Table 1).
16
Table 1. WISH demographic and baseline characteristics by treatment
Measurements
Placebo
(N=175)
ISP
(N=175)
p value
Age 60.7 (6.7) 60.8 (7.0) 0.50
Type of menopause
†
Natural
Surgical
154 (88.26)
20 (11.43)
160 (91.40)
15 (8.60)
0.36
Race
‡
White (non-Hispanic)
Black (non-Hispanic)
Hispanic
Asian or Pacific Islander
118 (67.43)
11 (6.29)
24 (13.71)
22 (12.57)
105 (60.00)
14 (8.00)
32 (18.29)
23 (13.14)
0.52
Years since menopause
<5 years
5-10 years
>10 years
Unknown
38 (21.71)
45 (25.71)
71 (40.57)
21 (12.00)
35 (20.00)
50 (28.57)
74 (52.29)
16 (9.14)
0.77
CIMT (mm)
<0.75
>=0.75
49 (28.0)
126 (72.0)
48 (27.5)
127 (72.5)
0.90
BMI (kg/m
2
) 26.65 (5.35) 26.48 (5.02) 0.76
Weight (lbs) 152.77 (32.45) 152.00 (29.74) 0.82
Waist Circumference (inches) 32.44 (4.87) 32.56 (4.27) 0.80
Hip Circumference (inches) 40.17 (4.75) 40.32 (3.68) 0.75
Waist/hip Ratio 0.81 (0.06) 0.81 (0.06) 0.92
Glucose (mg/dL)
‡
95.81 (10.06) 94.93 (8.92) 0.38
Total Cholesterol (mg/dL) 221.26 (33.98) 218.78 (30.33) 0.47
HDL-cholesterol (mg/dL) 63.70 (17.75) 62.71 (16.62) 0.59
LDL-cholesterol (mg/dL) 135.83 (32.19) 133.63 (27.11) 0.49
Triglycerides (mg/dL) 108.63 (59.44) 112.25 (67.48) 0.60
NOTE: Mean(SD) or n(%).
Group differences on categorized variables were calculated by chi square test, and group
differences on continuous variables were calculated by independent t test.
† One observation in placebo group was missing
‡ One observation in ISP group was missing
Among the ISP participants, 46.3% were non-equol producers, 20.0% were
intermittent equol producers, and 22.3% were consistent equol producers; equol producer
17
status was not identified on 11.4% of ISP subjects due to incomplete data on equol levels.
At baseline, there were no statistically significant differences in mean anthropometric and
metabolic measurements among the equol producer subgroups (Table 2).
Table 2. Baseline anthropometric and metabolic characteristics stratified by equol
producer subgroups
Variables
Placebo
N=175
ISP
P
Non-equol
producer
N=81
Intermittent
equol
producer
N=35
Consistent
equol
producer
N=39
BMI (kg/m
2
) 26.65 (0.40) 26.37 (0.58) 26.04 (0.89) 27.21 (0.84) 0.78
Weight (lbs) 152.77(2.89) 152.61 (3.51) 149.86 (5.34) 153.54 (5.06) 0.96
Waist
Circumference
(inches)
32.44 (0.35) 32.59 (0.51) 32.04 (0.78) 33.12 (0.74) 0.78
Hip
Circumference
(inches)
40.17 (0.34) 40.47 (0.50) 39.76 (0.76) 40.59 (0.72) 0.83
Waist-hip
Ratio
0.81 (0.005) 0.80 (0.006) 0.81 (0.010) 0.82 (0.010) 0.84
Glucose
(mg/dl)
95.81 (0.73) 95.51 (1.07) 94.57 (1.63) 94.97 (1.55) 0.89
Total
Cholesterol
(mg/dL)
221.26(2.43) 216.95 (3.56) 212.66 (5.42) 227.05 (5.14) 0.20
HDL-
cholesterol
(mg/dL)
63.70 (1.30) 62.25 (1.91) 62.00 (2.91) 62.08 (2.75) 0.88
LDL-
cholesterol
(mg/dL)
135.83(2.24) 131.43 (3.29) 130.67 (5.00) 140.53 (4.74) 0.33
Triglycerides
(mg/dL)
108.63(4.89) 116.36(7.18) 99.91 (10.93) 122.21(10.35
)
0.39
NOTE: Mean (SE), group differences were calculated using ANOVA.
In total 20 subjects in ISP group were not included in this analysis because of missing
information on equol-levels.
18
3.2 ANTHROPOMETRIC MEASUREMENTS
Tables 3 to 7 summarize the mean change from baseline in anthropometric
measurements in all subjects and subgroups. In all subjects, the mean change in BMI,
weight, hip circumference, waist circumference and waist to hip ratio in the ISP group
did not differ from the placebo group. Among the subgroups defined by time since
menopause, ethnicity and age at randomization, the treatment groups also did not differ
on mean change of these anthropometric measurements (Tables 3-7).
Table 3. Mean change in BMI (kg/m
2
) by treatment
N(placebo/ISP) Placebo ISP p
All subjects 350(175/175) 0.04 (0.11) 0.08 (0.10) 0.72
Time since menopause
<5 years 73 (38/35) 0.24 (0.15) 0.01 (0.17) 0.24
5-10 years 95 (45/50) 0.11 (0.16) 0.04 (0.14) 0.69
>10 years 145 (71/74) -0.16 (0.29) 0.04 (0.27) 0.26
Ethnicity
†
White (non-Hispanic) 223 (118/105) 0.13 (0.15) 0.12 (0.14) 0.90
Black (non-Hispanic) 25 (11/14) 0.60 (0.52) 0.49 (0.44) 0.85
Hispanic 56 (24/32) -0.52 (0.31) -0.03 (0.22) 0.16
Asian 45 (22/23) -0.12 (0.17) -0.12 (0.13) 0.99
Age at randomization
≤55 88 (46/42) 0.11 (0.16) 0.16 (0.16) 0.77
56-60 175 (84/91) -0.03 (0.16) -0.001(0.14) 0.86
>60 87 (45/42) 0.24 (0.14) 0.32 (0.14) 0.74
NOTE: Mean (SE) change from baseline, adjusted for randomization stratum.
Treatment groups compared using generalized estimating equations with identity link
function and exchangeable correlation structure.
† Ethnicity unknown for 1 participant and excluded from the model.
19
Table 4. Mean change in weight (lbs) by treatment
N(placebo/ISP) Placebo ISP p
All Subjects 350(175/175) 0.25 (0.63) 0.39 (0.57) 0.82
Time since menopause
<5 years 73 (38/35) 1.51 (0.89) 0.07 (0.99) 0.20
5-10 years 95 (45/50) 0.70 (0.93) 0.13 (0.79) 0.60
>10 years 145 (71/74) -1.10 (1.70) 0.05 (1.57) 0.27
Ethnicity
†
White (non-Hispanic) 223 (118/105) 0.71 (0.84) 0.56 (0.79) 0.82
Black (non-Hispanic) 25 (11/14) 4.07 (3.18) -3.02 (2.67) 0.77
Hispanic 56 (24/32) -2.98 (1.73) -0.16 (1.27) 0.16
Asian 45 (22/23) -0.62 (0.89) -0.64 (0.71) 0.98
Age at randomization
≤55 88 (46/42) 0.76 (0.96) 1.00 (0.94) 0.82
56-60 175 (84/91) -0.22 (0.92) -0.13 (0.78) 0.93
>60 87 (45/42) 1.55 (0.82) 1.88 (0.79) 0.79
NOTE: Mean (SE) change from baseline, adjusted for randomization stratum.
Treatment groups compared using generalized estimating equations with identity link
function and exchangeable correlation structure.
† Ethnicity unknown for 1 participant and excluded from the model.
Table 5. Mean change in hip circumference (inches) by treatment
N(placebo/ISP) Placebo ISP p
All Subjects 350(175/175) 0.22 (0.19) 0.15 (0.16) 0.71
Time since menopause
<5 years 73 (38/35) 0.61 (0.38) 0.05 (0.27) 0.17
5-10 years 95 (45/50) 0.26 (0.30) 0.20 (0.24) 0.83
>10 years 145 (71/74) -0.28 (0.36) -0.14 (0.34) 0.67
Ethnicity
†
White (non-Hispanic) 223 (118/105) 0.34 (0.22) 0.41 (0.21) 0.78
Black (non-Hispanic) 25 (11/14) -0.59 (0.58) 0.52 (0.58) 0.11
Hispanic 56 (24/32) 1.16 (0.90) 0.13 (0.46) 0.11
Asian 45 (22/23) -0.30 (0.27) -0.45 (0.21) 0.57
Age at randomization
≤55 88 (46/42) 0.48 (0.39) 0.20 (0.24) 0.38
56-60 175 (84/91) 0.12 (0.24) -0.02 (0.23) 0.66
>60 87 (45/42) 0.25 (0.43) 0.44 (0.43) 0.40
NOTE: Mean (SE) change from baseline, adjusted for randomization stratum.
Treatment groups compared using generalized estimating equations with identity link
function and exchangeable correlation structure.
† Ethnicity unknown for 1 participant and excluded from the model.
20
Table 6. Mean change in waist circumference (inches) by treatment
N(placebo/ISP) Placebo ISP p
All Subjects 350(175/175) -0.41 (0.18) -0.37 (0.18) 0.82
Time since menopause
<5 years 73 (38/35) -0.29 (0.29) -0.33 (0.35) 0.93
5-10 years 95 (45/50) -0.72 (0.27) -0.46 (0.24) 0.39
>10 years 145 (71/74) -0.27 (0.43) -0.18 (0.42) 0.77
Ethnicity
†
White (non-Hispanic) 223 (118/105) -0.39 (0.24) -0.44 (0.24) 0.84
Black (non-Hispanic) 25 (11/14) -1.08 (0.62) -1.00 (0.73) 0.92
Hispanic 56 (24/32) -0.35 (0.61) -0.02 (0.45) 0.49
Asian 45 (22/23) -0.33 (0.30) -0.47 (0.30) 0.64
Age at randomization
≤55 88 (46/42) -0.62 (0.29) -0.80 (0.35) 0.63
56-60 175 (84/91) -0.21 (0.24) -0.13 (0.21) 0.76
>60 87 (45/42) -0.16 (0.41) 0.02 (0.40) 0.65
NOTE: Mean (SE) change from baseline, adjusted for randomization stratum.
Treatment groups compared using generalized estimating equations with identity link
function and exchangeable correlation structure.
† Ethnicity unknown for 1 participant and excluded from the model.
Table 7. Mean change in waist to hip ratio by treatment
N(placebo/ISP) Placebo ISP p
All Subjects 350(175/175) 0.0050 (0.0049) 0.0053 (0.0044) 0.95
Time since menopause
<5 years 73 (38/35) -0.0076 (0.0107) 0.0073 (0.0085) 0.18
5-10 years 95 (45/50) 0.0135 (0.0074) 0.0064 (0.0061) 0.40
>10 years 145 (71/74) 0.0117 (0.0083) 0.0067 (0.0086) 0.52
Ethnicity
†
White (non-
Hispanic)
223 (118/105)
0.0024 (0.0051) 0.0016 (0.0056) 0.90
Black (non-
Hispanic)
25 (11/14)
0.0350 (0.0141) 0.0118 (0.0135) 0.13
Hispanic 56 (24/32) -0.0201 (0.0245) -0.0045 (0.0137) 0.32
Asian 45 (22/23) 0.0157 (0.0073) 0.0219 (0.0078) 0.41
Age at randomization
≤55 88 (46/42) 0.0036 (0.0100) 0.0148 (0.0082) 0.24
56-60 175 (84/91) 0.0030 (0.0056) 0.0030 (0.0051) 0.99
>60 87 (45/42) -0.0013 (0.0090) -0.0094 (0.0095) 0.41
Mean (SE) change from baseline, adjusted for randomization stratum.
Treatment groups compared using generalized estimating equations with identity link
function and exchangeable correlation structure.
† Ethnicity unknown for 1 participant and excluded from the model.
21
3.3 METABOLIC MEASUREMENTS
Tables 8 to 12 summarize the mean change in total cholesterol, HDL-cholesterol,
LDL-cholesterol, glucose and triglycerides in all subjects and subgroups. Among all
subjects, there was no statistically significant difference in the mean change of total
cholesterol between the ISP and placebo groups. However, in those who experienced
menopause more than 10 years ago, total cholesterol increased on average in the placebo
group but decreased on average in the ISP group (p=0.01). In Hispanic women, the mean
total cholesterol also decreased in the ISP group and increased in the placebo group
(p=0.02). Age at randomization did not modify treatment effect on total cholesterol levels
(Table 8).
Table 8. Mean change in total cholesterol (mg/dL) by treatment
N(placebo/ISP) Placebo ISP p
All Subjects 350(175/175) -0.32 (2.80) -2.29 (2.58) 0.44
Time since menopause
<5 years 73 (38/35) 2.11 (3.90) 5.51 (3.73) 0.46
5-10 years 95 (45/50) -11.26 (5.56) -3.63 (4.62) 0.15
>10 years 145 (71/74) 4.28 (3.99) -5.53 (4.27) 0.01
Ethnicity
†
White (non-Hispanic) 223 (118/105) -0.34 (2.97) 0.29 (2.89) 0.83
Black (non-Hispanic) 25 (11/14) 15.35 (9.34) 6.07 (6.54) 0.20
Hispanic 56 (24/32) 9.03 (8.62) -6.93 (5.62) 0.02
Asian 45 (22/23) -6.68 (8.00) -3.48 (9.33) 0.74
Age at randomization
≤55 88 (46/42) -2.06 (4.33) 2.16 (4.09) 0.37
56-60 175 (84/91) -0.74 (4.06) -4.70 (3.63) 0.29
>60 87 (45/42) 7.53 (4.99) 2.74 (5.04) 0.26
Mean (SE) change from baseline, adjusted for randomization stratum.
Treatment groups compared using generalized estimating equations with identity link
function and exchangeable correlation structure.
† Ethnicity unknown for 1 participant and excluded from the model.
Average HDL-cholesterol increased both in the placebo and ISP groups. Among all
subjects and women younger than 55 years old, the mean change in HDL-cholesterol was
higher in the ISP group than in the placebo group (p=0.05). In white women and women
22
who experienced menopause less than 5 years ago, the mean change in the ISP group was
significantly higher than in the placebo group (both p=0.04) (Table 9).
Table 9. Mean change in HDL-cholesterol (mg/dL) by treatment
N(placebo/ISP) Placebo ISP p
All Subjects 350(175/175) 1.09 (0.67) 2.60 (0.74) 0.05
Time since menopause
<5 years 73 (38/35) 0.17 (1.20) 3.82 (1.06) 0.04
5-10 years 95 (45/50) 0.56 (1.03) 2.41 (1.35) 0.14
>10 years 145 (71/74) 1.80 (1.53) 2.52 (1.63) 0.55
Ethnicity
†
White (non-Hispanic) 223 (118/105) 1.06 (0.90) 3.13 (1.07) 0.04
Black (non-Hispanic) 25 (11/14) 0.67 (2.91) 1.32 (2.98) 0.84
Hispanic 56 (24/32) 2.20 (1.75) 0.57 (1.39) 0.30
Asian 45 (22/23) 0.99 (1.36) 5.01 (1.68) 0.07
Age at randomization
≤55 88 (46/42) 0.28 (1.09) 3.44 (1.12) 0.05
56-60 175 (84/91) 1.25 (0.91) 2.29 (1.08) 0.30
>60 87 (45/42) 2.10 (1.98) 2.61 (1.84) 0.77
Mean (SE) change from baseline, adjusted for randomization stratum.
Treatment groups compared using generalized estimating equations with identity link
function and exchangeable correlation structure.
† Ethnicity unknown for 1 participant and excluded from the model.
In all subjects, the mean change in LDL-cholesterol did not differ between the
placebo and ISP groups. However, among women who experienced menopause more
than 10 years ago, LDL-cholesterol decreased significantly in the ISP group compared to
the placebo group (p=0.01). In Hispanic women, LDL-cholesterol also significantly
decreased (p=0.04). Age at randomization did not modify the treatment effect on LDL-
cholesterol levels (Table 10).
23
Table 10. Mean change in LDL-cholesterol (mg/dL) by treatment
N(placebo/ISP) Placebo ISP p
All Subjects 350(175/175) -2.81 (2.58) -5.02 (2.41) 0.34
Time since menopause
<5 years 73 (38/35) -0.14 (3.43) -0.37 (3.43) 0.90
5-10 years 95 (45/50) -11.93 (4.98) -5.35 (4.40) 0.17
>10 years 145 (71/74) 2.15 (4.23) -7.25 (4.33) 0.01
Ethnicity
†
White (non-Hispanic) 223 (118/105) -3.75 (2.76) -3.38 (2.72) 0.89
Black (non-Hispanic) 25 (11/14) 16.07 (8.44) 3.43 (5.92) 0.07
Hispanic 56 (24/32) 6.21 (8.47) -7.93 (5.39) 0.04
Asian 45 (22/23) -7.47 (7.12) -6.00 (8.43) 0.86
Age at randomization
≤55 88 (46/42) -3.52 (4.22) -2.83 (3.50) 0.87
56-60 175 (84/91) -3.38 (3.66) -6.05 (3.55) 0.44
>60 87 (45/42) 2.78 (4.76) -1.74 (4.69) 0.29
Mean (SE) change from baseline, adjusted for randomization stratum.
Treatment groups compared using generalized estimating equations with identity link
function and exchangeable correlation structure.
† Ethnicity unknown for 1 participant and excluded from the model.
In all subjects, triglycerides increased both in the ISP and placebo groups; the
increase in the ISP group was marginally significantly lower than in the placebo group
(p=0.05). In white women, the mean increase in the ISP group was also statistically
significantly lower than in the placebo group (p=0.03). In women aged 56 to 60 years old,
the triglycerides level was reduced in the ISP group, relative to the placebo group
(p=0.01). Time since menopause did not modify the treatment effect on triglycerides
levels (Table 11).
24
Table 11. Mean change in triglycerides (mg/dL) by treatment
N(placebo/ISP) Placebo ISP P
All Subjects 350(175/175) 10.31 (5.91) 2.25 (4.47) 0.05
Time since menopause
<5 years 73 (38/35) 14.09 (6.04) 6.34 (8.26) 0.38
5-10 years 95 (45/50) 0.01 (6.47) -4.05 (4.42) 0.50
>10 years 145 (71/74) 1.60 (6.53) -4.10 (7.29) 0.25
Ethnicity
†
White (non-Hispanic) 223 (118/105) 15.80 (8.51) 4.96 (6.22) 0.03
Black (non-Hispanic) 25 (11/14) -7.05 (10.53) 11.28 (8.36) 0.17
Hispanic 56 (24/32) 3.11 (12.14) 2.59 (11.73) 0.96
Asian 45 (22/23) 3.53 (11.09) -10.35 (7.89) 0.19
Age at randomization
≤55 88 (46/42) 16.11 (14.29) 10.70 (9.16) 0.63
56-60 175 (84/91) 6.74 (5.18) -5.06 (4.36) 0.01
>60 87 (45/42) 12.80 (6.36) 8.93 (6.21) 0.53
Mean (SE) change from baseline, adjusted for randomization stratum.
Treatment groups compared using generalized estimating equations with identity link
function and exchangeable correlation structure.
† Ethnicity unknown for 1 participant and excluded from the model.
The treatment groups did not differ on the mean change of glucose. In black women
the mean decrease in the ISP group was marginally significantly less than the placebo
group (p=0.05). The treatment effect on glucose was not modified by time since
menopause or age (Table 12).
25
Table 12. Mean change in glucose (mg/dL) by treatment
N(placebo/ISP) Placebo ISP p
All Subjects 350(175/175) -2.68 (0.90) -0.81 (0.95) 0.12
Time since menopause
<5 years 73 (38/35) -4.55 (2.90) 1.99 (2.63) 0.20
5-10 years 95 (45/50) -1.71 (1.26) -2.12 (1.59) 0.76
>10 years 145 (71/74) -2.80 (1.20) -1.54 (1.32) 0.33
Ethnicity
†
White (non-Hispanic) 223 (118/105) -3.10 (1.11) -1.06 (1.27) 0.25
Black (non-Hispanic) 25 (11/14) -1.74 (2.31) 3.39 (2.04) 0.05
Hispanic 56 (24/32) -4.94 (2.92) -2.80 (1.58) 0.43
Asian 45 (22/23) -0.65 (1.83) 0.35 (2.96) 0.67
Age at randomization
≤55 88 (46/42) -1.26 (1.39) 0.19 (1.27) 0.32
56-60 175 (84/91) -4.29 (1.26) -1.19 (1.46) 0.11
>60 87 (45/42) 0.09 (1.40) -0.65 (1.28) 0.62
Mean (SE) change from baseline, adjusted for randomization stratum.
Treatment groups compared using generalized estimating equations with identity link
function and exchangeable correlation structure.
† Ethnicity unknown for 1 participant and excluded from the model
3.4 EQUOL
Treatment effects on anthropometric measurements did not differ among equol
subgroups. However, there was evidence that equol status modifies treatment effects on
lipids. Compared to the mean change in the placebo group, the mean change of HDL-
cholesterol in the non-equol producer subgroup was marginally higher by comparison of
placebo (p=0.05, p
total
=0.17). Treatment effects on triglycerides significantly differed
among equol subgroups (p
total
=0.03). The mean change of triglycerides in the non-equol
producer subgroup and the consistent equol producer subgroup were significantly lower
than the placebo group (both p=0.04) (Table 13).
26
Table 13. Mean change in anthropometric and metabolic variables by equol producer
subgroup
Variables Estimate SE p p
total
†
BMI 0.80
Placebo group Reference -- --
Non equol producer 0.09 0.12 0.47
Intermittent equol producer 0.04 0.17 0.80
Consistent equol producer -0.07 0.17 0.69
Waist circumference 0.47
Placebo group Reference -- --
Non equol producer -0.15 0.22 0.48
Intermittent equol producer 0.25 0.33 0.45
Consistent equol producer 0.25 0.27 0.36
Weight 0.80
Placebo group Reference -- --
Non equol producer 0.45 0.71 0.53
Intermittent equol producer 0.15 1.00 0.88
Consistent equol producer -0.49 0.97 0.61
Hip circumference 0.82
Placebo group Reference -- --
Non equol producer -0.03 0.24 0.89
Intermittent equol producer 0.02 0.31 0.95
Consistent equol producer -0.23 0.25 0.36
Waist/hip Ratio 0.63
Placebo group Reference -- --
Non equol producer 0.00047 0.0054 0.38
Intermittent equol producer -0.0071 0.0106 0.50
Consistent equol producer -0.0021 0.0067 0.75
Glucose 0.29
Placebo group Reference -- --
Non equol producer 0.83 0.96 0.39
Intermittent equol producer -0.57 1.23 0.64
Consistent equol producer 6.18 3.61 0.09
Total Cholesterol 0.31
Placebo group Reference -- --
Non equol producer -0.79 3.07 0.80
Intermittent equol producer 0.91 3.50 0.79
Consistent equol producer -7.23 4.03 0.07
HDL-cholesterol 0.17
Placebo group Reference -- --
Non equol producer 1.73 0.89 0.05
Intermittent equol producer 0.64 1.24 0.61
Consistent equol producer 2.08 1.22 0.09
LDL-cholesterol 0.34
Placebo group Reference -- --
27
Table 13. Mean change in anthropometric and metabolic variables by equol producer
subgroup, continued
Non equol producer -0.92 2.75 0.74
Intermittent equao producer -0.51 3.27 0.88
Consistent equol producer -7.07 3.79 0.06
Triglycerides 0.03
Placebo group Reference -- --
Non equol producer -10.03 4.91 0.04
Intermittent equol producer 2.42 5.05 0.63
Consistent equol producer -12.64 6.06 0.04
NOTE: Calculated by generalized estimating equations, adjusted for randomization
stratum, with identity link function and exchangeable correlation structure.
† Overall p value for differences among groups
3.5 PLASMA IFLS
Tables 14 to 23 summarize the association of plasma IFLs on mean change of
anthropometric and metabolic variables. Both model 1 and model 2 indicate that plasma
isoflavone levels were not statistically significantly associated with plasma lipids. After
adjustment for randomization stratum, plasma O-desmethylangolensin level was
positively associated with HDL cholesterol (β=1.84; p=0.03), however, after adding
treatment for adjustment, the association was no longer significant (β=1.44; p=0.09)
(Table 14). In both models, plasma dihydrogenistein was negatively associated with hip
circumference (β=-0.56; p=0.01) but positively associated with waist circumference
(β=0.53; p=0.02). Plasma genistein level was only negatively associated with hip
circumference (β=-0.16; p=0.01) (Tables 22, 23).
28
Table 14. association among plasma IFLs and change in HDL-cholesterol
Model 1 Model 2
Estimate (SE) p Estimate (SE) p
Daidzein 0.06 (0.47) 0.90 0.31 (0.46) 0.50
Dihydrodaidzein 0.55 (0.96) 0.55 0.87 (0.98) 0.33
O-desmethylangolensin 1.44 (0.82) 0.09 1.84 (0.81) 0.03
Equol 1.07 (0.93) 0.19 1.32 (1.02) 0.12
Genistein -0.66 (0.38) 0.10 -0.39 (0.37) 0.32
O-glycitein -0.42 (8.50) 0.96 1.21 (8.04) 0.88
Dihydrogenistein -1.17 (1.23) 0.40 -0.95 (1.23) 0.49
NOTE: Mean(E)*1000
Analyzed by generalized estimating equations with identity link function and
exchangeable correlation structure.
Model 1 was adjusted for treatment and randomization stratum, model 2 was adjusted for
randomization stratum
Table 15. Association among plasma IFLs and change in total cholesterol
Model 1 Model 2
Estimate (SE) p Estimate (SE) P
Daidzein 0.41 (1.58) 0.79 0.05 (1.55) 0.97
Dihydrodaidzein -1.19 (2.79) 0.69 -1.63 (2.77) 0.59
O-desmethylangolensin 3.01 (2.33) 0.22 2.25 (2.36) 0.36
Equol 4.02 (4.06) 0.15 3.61 (3.84) 0.19
Genistein -1.17 (1.47) 0.44 -1.41 (1.41) 0.33
O-glycitein -10.76 (23.76) 0.66 -12.96 (23.68) 0.59
Dihydrogenistein -2.17 (4.04) 0.64 -2.48 (4.00) 0.59
NOTE: Mean(E)*1000
Analyzed by generalized estimating equations with identity link function and
exchangeable correlation structure.
Model 1 was adjusted for treatment and randomization stratum, model 2 was adjusted for
randomization stratum
29
Table 16. Association among plasma IFLs and change in BMI
Model 1 Model 2
Estimate (SE) p Estimate (SE) P
Daidzein 0.04(0.07) 0.52 0.05 (0.06) 0.44
Dihydrodaidzein -0.06 (0.11) 0.62 -0.05 (0.11) 0.68
O-desmethylangolensin 0.14 (0.13) 0.24 0.15 (0.13) 0.21
Equol -0.15 (0.08) 0.23 -0.14 (0.08) 0.25
Genistein 0.02 (0.05) 0.59 0.03 (0.05) 0.49
O-glycitein -1.12 (0.72) 0.23 -1.06 (0.72) 0.25
Dihydrogenistein -0.42 (0.18) 0.09 -0.42 (0.18) 0.10
NOTE: Mean(E)*1000
Analyzed by generalized estimating equations with identity link function and
exchangeable correlation structure.
Model 1 was adjusted for treatment and randomization stratum, model 2 was adjusted for
randomization stratum
Table 17. Association among plasma IFLs and change in LDL-cholesterol
Model 1 Model 2
Estimate (SE) p Estimate (SE) P
Daidzein -0.45 (1.33) 0.74 -0.82 (1.30) 0.54
Dihydrodaidzein -2.88 (2.30) 0.30 -3.36 (2.29) 0.23
O-desmethylangolensin 1.05 (1.87) 0.58 0.28 (1.91) 0.88
Equol 3.61 (3.85) 0.21 3.15 (3.64) 0.27
Genistein -1.51 (1.23) 0.25 -1.77 (1.18) 0.16
O-glycitein -9.99 (21.15) 0.65 -12.61 (21.09) 0.56
Dihydrogenistein -3.74 (3.65) 0.41 -4.09 (3.60) 0.37
NOTE: Mean(E)*1000
Analyzed by generalized estimating equations with identity link function and
exchangeable correlation structure.
Model 1 was adjusted for treatment and randomization stratum, model 2 was adjusted for
randomization stratum
30
Table 18. Association among plasma IFLs and change in triglycerides
Model 1 Model 2
Estimate (SE) p Estimate (SE) p
Daidzein 4.54 (2.92) 0.13 1.79 (2.75) 0.52
Dihydrodaidzein 4.94 (5.41) 0.39 1.43 (5.56) 0.80
O-desmethylangolensin 3.47 (3.95) 0.39 -1.21 (3.81) 0.75
Equol -4.13 (4.72) 0.40 -6.47 (5.22) 0.21
Genistein 4.76 (2.37) 0.07 2.14 (2.29) 0.37
O-glycitein 7.42 (41.81) 0.85 -9.36 (36.88) 0.81
Dihydrogenistein 13.38 (8.94) 0.17 10.86 (8.80) 0.26
NOTE: Mean(E)*1000
Analyzed by generalized estimating equations with identity link function and
exchangeable correlation structure.
Model 1 was adjusted for treatment and randomization stratum, model 2 was adjusted for
randomization stratum.
Table 19. Association among plasma IFLs and change in glucose
Model 1 Model 2
Estimate (SE) P Estimate (SE) P
Daidzein -0.67 (1.07) 0.53 0.11 (0.85) 0.89
Dihydrodaidzein -3.51 (4.55) 0.46 -1.59 (3.82) 0.69
O-desmethylangolensin 1.42 (2.16) 0.52 2.20 (2.27) 0.34
Equol 3.50 (3.13) 0.17 3.83 (3.32) 0.14
Genistein -0.96 (0.86) 0.26 -0.22 (0.71) 0.75
O-glycitein 8.16 (28.36) 0.51 12.85 (14.18) 0.33
Dihydrogenistein 3.14 (2.23) 0.18 4.27 (2.51) 0.08
NOTE: Mean(E)*1000
Analyzed by generalized estimating equations with identity link function and
exchangeable correlation structure.
Model 1 was adjusted for treatment and randomization stratum, model 2 was adjusted for
randomization stratum.
31
Table 20. Association among plasma IFLs and change in weight
Model 1 Model 2
Estimate (SE) p Estimate (SE) P
Daidzein 0.19 (0.37) 0.60 0.21 (0.36) 0.53
Dihydrodaidzein -0.41 (0.59) 0.56 -0.36 (0.60) 0.61
O-desmethylangolensin 0.73 (0.70) 0.26 0.77 (0.69) 0.23
Equol -0.86 (0.45) 0.23 -0.81 (0.45) 0.24
Genistein 0.11 (0.27) 0.66 0.14 (0.26) 0.58
O-glycitein -7.04 (3.96) 0.18 -6.79 (3.96) 0.19
Dihydrogenistein -2.48 (1.05) 0.09 -2.45 (1.05) 0.09
NOTE: Mean(E)*1000
Analyzed by generalized estimating equations with identity link function and
exchangeable correlation structure.
Model 1 was adjusted for treatment and randomization stratum, model 2 was adjusted for
randomization stratum.
Table 21. Association among plasma IFLs and change in waist to hip ratio
Model 1 Model 2
Estimate (SE) p Estimate (SE) P
Daidzein 0.0018 (0.002) 0.33 0.0017 (0.0019) 0.35
Dihydrodaidzein -0.0033 (0.0037) 0.35 -0.0034 (0.0036) 0.34
O-desmethylangolensin -0.0027 (0.0035) 0.42 -0.0027 (0.0034) 0.42
Equol -0.0007 (0.003) 0.82 -0.0008 (0.003) 0.80
Genistein 0.0031 (0.0017) 0.07 0.0029 (0.0017) 0.08
O-glycitein 0.020 (0.0305) 0.48 0.0195 (0.0303) 0.49
Dihydrogenistein 0.0004 (0.0038) 0.92 0.0003 (0.0038) 0.93
NOTE: Mean(E)*1000
Analyzed by generalized estimating equations with identity link function and
exchangeable correlation structure.
Model 1 was adjusted for treatment and randomization stratum, model 2 was adjusted for
randomization stratum.
32
Table 22. Association among plasma IFLs and change in hip circumference
Model 1 Model 2
Estimate (SE) p Estimate (SE) P
Daidzein -0.08 (0.08) 0.31 -0.09 (0.08) 0.28
Dihydrodaidzein -0.05 (0.13) 0.72 -0.06 (0.13) 0.67
O-desmethylangolensin 0.18 (0.16) 0.29 0.16 (0.16) 0.33
Equol -0.16 (0.16) 0.29 -0.17 (0.17) 0.27
Genistein -0.16 (0.06) 0.01 -0.16 (0.06) 0.01
O-glycitein -0.94 (1.17) 0.39 -0.98 (1.18) 0.37
Dihydrogenistein -0.56 (0.23) 0.01 -0.56 (0.23) 0.01
NOTE: Mean(E)*1000
Analyzed by generalized estimating equations with identity link function and
exchangeable correlation structure.
Model 1 was adjusted for treatment and randomization stratum, model 2 was adjusted for
randomization stratum.
Table 23. Association among plasma IFLs and change in waist circumference
Model 1 Model 2
Estimate (SE) p Estimate (SE) P
Daidzein -0.01 (0.08) 0.93 0.02 (0.08) 0.81
Dihydrodaidzein 0.22 (0.17) 0.26 0.23 (0.17) 0.23
O-desmethylangolensin -0.02 (0.16) 0.90 0.00 (0.16) 0.98
Equol 0.22 (0.11) 0.14 0.23 (0.11) 0.12
Genistein 0.02 (0.07) 0.78 0.03 (0.06) 0.66
O-glycitein 0.38 (1.37) 0.79 0.46 (1.34) 0.74
Dihydrogenistein 0.52 (0.25) 0.02 0.53 (0.25) 0.02
NOTE: Mean(E)*1000
Analyzed by generalized estimating equations with identity link function and
exchangeable correlation structure.
Model 1 was adjusted for treatment and randomization stratum, model 2 was adjusted for
randomization stratum.
33
CHAPTER 4: DISCUSSION
WISH is one of the largest randomized clinical trials evaluating the health effects of
soy isoflavones in postmenopausal women. This analysis indicates that 25g soy protein
containing 85mg isoflavones daily has no overall effect on anthropometric measurements,
including BMI, weight, hip circumference, waist circumference and waist to hip ratio in
postmenopausal women. Previous observational studies have evaluated dietary isoflavone
intake in relation to anthropometric measurements. Differing from our results, one study
reported that genistein intake was negatively associated with waist circumference and
BMI. These differences are likely primarily due to the differing designs (randomized vs.
observational study, with uncontrolled confounding biases) as well as possible
measurement errors in assessment of isoflavone intake using a self-administered food
frequency questionnaire (Goodman-Gruen and Kritz-Silverstein 2001; Goodman-Gruen
and Kritz-Silverstein 2003). This analysis is the first study to examine plasma isoflavone
concentrations in relation to anthropometric and metabolic measurements. In analyses
relating plasma isoflavone levels to anthropometric measures, we noted some
associations of small magnitude of plasma dihydrogenistein and genistein with hip and
waist circumference. Studies in rodents suggested that high levels of isoflavones may
decrease adipose deposition by increasing fatty acid oxidation, LDL receptor activity and
mitochondrial enzyme activities. However, the mechanism in human is not clear.
In our analysis, we observed increases in HDL-cholesterol; although the increase
was marginally significant in the total ISP participants relative to placebo (p=0.05), the
34
treatment group difference was significant in white women and women who experienced
menopause less than 5 years ago (p=0.04). We also observed that ISP attenuated the
increase in triglycerides observed in both treatment groups over trial follow-up. This
effect was marginally significant in all participants (p=0.05) but was significant in white
women (p=0.03). While total cholesterol and LDL cholesterol concentrations did not
change in the total ISP participants relative to placebo, we did observe that in Hispanic
women (p=0.02) and those who experienced menopause more than 10 years ago (p=0.01),
total cholesterol and LDL-cholesterol decreased in ISP-treated compared to placebo-
treated women. Plasma isoflavone concentrations were not significantly associated with
lipid levels. Previous studies have reported inconsistent findings regarding associations of
isoflavones with lipids. Our analysis is consistent with a recent study, which suggested
that 30gm daily of soy protein is positively associated with HDL-cholesterol but
negatively associated with serum cholesterol, triglycerides, and LDL-cholesterol levels
(H K Jassi 2010).
To date, there is no standard method to define equol-producing capacity. Since
equol-producing capability varies among individuals and over time within individuals,
we categorized equol status into three subgroups according to plasma equol levels
measured repeatedly over the trial – non-equol producer, intermittent, and consistent
equol producers. Our identification standard is more complex than previous studies which
had only two categories (non-equol producer and equol producer), as other studies were
cross-sectional in design. Soy isoflavones increased HDL-cholesterol more effectively
among non-equol producers compared to placebo, although the effect was marginally
significant (p=0.05). Equol status also modified the association of isoflavone treatment
35
with triglycerides; triglycerides decreased significantly by ISP treatment in both non-
equol producers and consistent equol producers, but not among intermittent equol
producers. Equol-producing capability was not associated with anthropometric
measurements. Different from Meyer’s findings, which reported that total cholesterol,
LDL-cholesterol, and plasma triglycerides reduced in equol producers (Meyer, Larkin et
al. 2004), we did not find differential effect of isoflavone supplementation of equol
producing capability on LDL-cholesterol or total cholesterol. This difference may be due
to the difference in determining equol status. The relatively small number of equol
producers in the other study may be another explanation for the difference.
The typical consumption of soy protein among Asian populations is approximately
30-50g per day, while other populations consume less than 5g per day soy protein in the
United States(Coward L 1993). Since soy protein with naturally occurring isoflavones,
but not isoflavone extract, carries a US Food and Drug Administration approval of a
health claim, and it appears that soy protein and isoflavones together are needed for
maximal activity, we chose soy protein as our trial study product. In this trial, the
isoflavone profile is similar to the reported intakes in typical Asian diets.
Our data have some strengths and limitations. In this randomized placebo controlled
study, we were able to address the influence of soy isoflavones in postmenopausal
women. Using the plasma isoflavones concentrations measured throughout the trial, we
were able to evaluate the influence of plasma IFLs and equol producing capability in
postmenopausal women. Moreover, repeated measurements of both isoflavone levels and
outcomes (anthropometrics and lipids) strengthened our analyses by using more
representative measurements. With a larger sample size, high compliance and longer
36
duration, we were able to have more power for detecting treatment effects. The major
limitation of this ancillary study is that the study hypothesis in our analysis was not
directly designed in WISH, thus generalization of the study results may be limited.
37
BIBLIOGRAPHY
Alicia A Thorp, P. R. H., Trevor A Mori, Alison M Coates, Jonathan D Buckley,
Jonathan Hodgson, Jackie Mansour, and Barbara J Meyer (2008). "Soy food
consumption does not lower LDL cholesterol in either equol or nonequol
producers." The American Journal of Clinical Nutrition 88(2).
Allen, J. K., D. M. Becker, et al. (2007). "Effect of soy protein-containing isoflavones on
lipoproteins in postmenopausal women." Menopause 14(1): 106-114.
Aoyama, T., K. Fukui, et al. (2000). "Soy protein isolate and its hydrolysate reduce body
fat of dietary obese rats and genetically obese mice (yellow KK)." Nutrition 16(5):
349-354.
Association, A. M. (2002). "Long-term use of estrogen-only hormone replacement
therapy (HRT) linked with increased risk of ovarian cancer." Ginecol Obstet Mex
70: 409-410.
Beral, V., D. Bull, et al. (2007). "Ovarian cancer and hormone replacement therapy in the
Million Women Study." Lancet 369(9574): 1703-1710.
Bolego, C., A. Poli, et al. (2003). "Phytoestrogens: pharmacological and therapeutic
perspectives." Curr Drug Targets 4(1): 77-87.
Cassidy, A. (2006). "Factors affecting the bioavailability of soy isoflavones in humans." J
AOAC Int 89(4): 1182-1188.
Coward L, B. N., Setchell KDR, Barnes S. (1993). "The isoflavones genistein and
daidzein in soy based foods from American and Asian diets. ." J Agricul Food Sci
41.
Dang, Z. and C. W. Lowik (2004). "The balance between concurrent activation of ERs
and PPARs determines daidzein-induced osteogenesis and adipogenesis." J Bone
Miner Res 19(5): 853-861.
38
de Lemos, M. L. (2001). "Effects of soy phytoestrogens genistein and daidzein on breast
cancer growth." Ann Pharmacother 35(9): 1118-1121.
Erberich, L. C., V. M. Alcantara, et al. (2002). "Hormone replacement therapy in
postmenopausal women and its effects on plasma lipid levels." Clin Chem Lab
Med 40(5): 446-451.
Fritsche S, S. H. (1999). "Occurrence of hormonally active compounds in food; a
review." Eur Food Res Technol 9: 153-179.
Genazzani, A. R. and M. Gambacciani (2006). "Effect of climacteric transition and
hormone replacement therapy on body weight and body fat distribution." Gynecol
Endocrinol 22(3): 145-150.
Goodman-Gruen, D. and D. Kritz-Silverstein (2001). "Usual dietary isoflavone intake is
associated with cardiovascular disease risk factors in postmenopausal women." J
Nutr 131(4): 1202-1206.
Goodman-Gruen, D. and D. Kritz-Silverstein (2003). "Usual dietary isoflavone intake
and body composition in postmenopausal women." Menopause 10(5): 427-432.
Gu, L., S. E. House, et al. (2006). "Metabolic phenotype of isoflavones differ among
female rats, pigs, monkeys, and women." J Nutr 136(5): 1215-1221.
H K Jassi, A. J., S Arora, R Chitra (2010). "EFFECT OF SOY PROTEINS Vs SOY
ISOFLAVONES ON LIPID PROFILE IN POSTMENOPAUSAL WOMEN."
Indian Journal of Clinical biochemistry 25(2): 201.
Hall, W. L., G. Rimbach, et al. (2005). "Isoflavones and endothelial function." Nutr Res
Rev 18(1): 130-144.
Hall, W. L., K. Vafeiadou, et al. (2006). "Soy-isoflavone-enriched foods and markers of
lipid and glucose metabolism in postmenopausal women: interactions with
genotype and equol production." Am J Clin Nutr 83(3): 592-600.
Ingram, D., K. Sanders, et al. (1997). "Case-control study of phyto-oestrogens and breast
cancer." Lancet 350(9083): 990-994.
39
Kenneth D. R, S., Sidney J. Cole (2006). "Method of Defining Equol-Producer Status and
Its Frequency among Vegetarians." The journal of Nutrition 136(8).
Kim, H. K., C. Nelson-Dooley, et al. (2006). "Genistein decreases food intake, body
weight, and fat pad weight and causes adipose tissue apoptosis in ovariectomized
female mice." J Nutr 136(2): 409-414.
Kirk, E. A., P. Sutherland, et al. (1998). "Dietary isoflavones reduce plasma cholesterol
and atherosclerosis in C57BL/6 mice but not LDL receptor-deficient mice." J
Nutr 128(6): 954-959.
Kreijkamp-Kaspers, S., L. Kok, et al. (2004). "Effect of soy protein containing
isoflavones on cognitive function, bone mineral density, and plasma lipids in
postmenopausal women: a randomized controlled trial." JAMA 292(1): 65-74.
Kuiper, G. G., E. Enmark, et al. (1996). "Cloning of a novel receptor expressed in rat
prostate and ovary." Proc Natl Acad Sci U S A 93(12): 5925-5930.
Llaneza, P., C. Gonzalez, et al. (2010). "Soy isoflavones, diet and physical exercise
modify serum cytokines in healthy obese postmenopausal women."
Phytomedicine.
Meli, R., M. Pacilio, et al. (2004). "Estrogen and raloxifene modulate leptin and its
receptor in hypothalamus and adipose tissue from ovariectomized rats."
Endocrinology 145(7): 3115-3121.
Meyer, B. J., T. A. Larkin, et al. (2004). "Limited lipid-lowering effects of regular
consumption of whole soybean foods." Ann Nutr Metab 48(2): 67-78.
Mullen, E., R. M. Brown, et al. (2004). "Soy isoflavones affect sterol regulatory element
binding proteins (SREBPs) and SREBP-regulated genes in HepG2 cells." J Nutr
134(11): 2942-2947.
Na, X. L., J. Ezaki, et al. (2008). "Isoflavone regulates lipid metabolism via expression of
related genes in OVX rats fed on a high-fat diet." Biomed Environ Sci 21(5): 357-
364.
40
Naaz, A., S. Yellayi, et al. (2003). "The soy isoflavone genistein decreases adipose
deposition in mice." Endocrinology 144(8): 3315-3320.
Niculescu, M. D., E. A. Pop, et al. (2007). "Dietary isoflavones differentially induce gene
expression changes in lymphocytes from postmenopausal women who form equol
as compared with those who do not." J Nutr Biochem 18(6): 380-390.
Nikander, E., A. Tiitinen, et al. (2004). "Effects of isolated isoflavonoids on lipids,
lipoproteins, insulin sensitivity, and ghrelin in postmenopausal women." J Clin
Endocrinol Metab 89(7): 3567-3572.
Owen, A. J., P. D. Roach, et al. (2004). "Regulation of low-density lipoprotein receptor
activity by estrogens and phytoestrogens in a HepG2 cell model." Ann Nutr
Metab 48(4): 269-275.
Potter, S. M., J. A. Baum, et al. (1998). "Soy protein and isoflavones: their effects on
blood lipids and bone density in postmenopausal women." Am J Clin Nutr 68(6
Suppl): 1375S-1379S.
Setchell, K. D., C. Clerici, et al. (2005). "S-equol, a potent ligand for estrogen receptor
beta, is the exclusive enantiomeric form of the soy isoflavone metabolite
produced by human intestinal bacterial flora." Am J Clin Nutr 81(5): 1072-1079.
Shields, T. S., N. S. Weiss, et al. (1999). "The additional risk of endometrial cancer
associated with unopposed estrogen use in women with other risk factors."
Epidemiology 10(6): 733-738.
Sudhaa Sharma, R. B., Vishal. R. Tandon, Annil Mahajan (2008). "Postmenopausal
Obesity." JK Science 10(3).
Uesugi, T., Y. Fukui, et al. (2002). "Beneficial effects of soybean isoflavone
supplementation on bone metabolism and serum lipids in postmenopausal
japanese women: a four-week study." J Am Coll Nutr 21(2): 97-102.
Wade, G. N., J. M. Gray, et al. (1985). "Gonadal influences on adiposity." Int J Obes 9
Suppl 1: 83-92.
41
Wagner, J. D., D. C. Schwenke, et al. (2003). "Soy protein with isoflavones, but not an
isoflavone-rich supplement, improves arterial low-density lipoprotein metabolism
and atherogenesis." Arterioscler Thromb Vasc Biol 23(12): 2241-2246.
Wangen, K. E., A. M. Duncan, et al. (2001). "Soy isoflavones improve plasma lipids in
normocholesterolemic and mildly hypercholesterolemic postmenopausal women."
Am J Clin Nutr 73(2): 225-231.
Washburn, S., G. L. Burke, et al. (1999). "Effect of soy protein supplementation on
serum lipoproteins, blood pressure, and menopausal symptoms in perimenopausal
women." Menopause 6(1): 7-13.
Wu, J., J. Oka, et al. (2007). "Possible role of equol status in the effects of isoflavone on
bone and fat mass in postmenopausal Japanese women: a double-blind,
randomized, controlled trial." Menopause 14(5): 866-874.
Yang, J. Y., S. J. Lee, et al. (2006). "Effect of genistein with carnitine administration on
lipid parameters and obesity in C57Bl/6J mice fed a high-fat diet." J Med Food
9(4): 459-467.
42
APPENDIX
Tables 24 to 33 show the association between change in IFLs and variables. In this
analysis the predictors were the mean change in plasma IFLs. The models showed similar
results.
Table 24. Association between change in plasma IFLs and change in HDL-cholesterol
Model 1 Model 2
Estimate (SE) P Estimate (SE) p
Daidzein 0.07 (0.48) 0.89 0.33 (0.46) 0.48
Dihydrodaidzein 0.54 (0.97) 0.56 0.90 (0.98) 0.31
O-desmethylangolensin 1.48 (0.81) 0.08 1.87 (0.80) 0.02
Equol 1.10 (0.94) 0.18 1.36 (1.03) 0.11
Genistein -0.70 (0.39) 0.09 -0.41 (0.38) 0.30
O-glycitein -2.18 (8.52) 0.79 -0.58 (8.00) 0.94
Dihydrogenistein -1.30 (1.23) 0.35 -1.06 (1.22) 0.44
NOTE: Mean(E)*1000
Analyzed by generalized estimating equations with identity link function and
exchangeable correlation structure.
Model 1 was adjusted for treatment and randomization stratum, model 2 was adjusted for
randomization stratum
Table 25. Association between change in plasma IFLs and change in total cholesterol
Model 1 Model 2
Estimate (SE) P Estimate (SE) p
Daidzein -0.43 (1.50) 0.78 -0.76 (1.48) 0.61
Dihydrodaidzein -2.49 (2.55) 0.39 -2.94 (2.57) 0.31
O-desmethylangolensin 2.51 (2.31) 0.30 1.79 (2.36) 0.47
Equol 3.99 (4.06) 0.16 3.56 (3.83) 0.20
Genistein -1.67 (1.49) 0.28 -1.87 (1.43) 0.20
O-glycitein -24.60 (22.59) 0.28 -26.56 (22.73) 0.24
Dihydrogenistein -2.32 (4.02) 0.61 -2.66 (3.98) 0.56
NOTE: Mean(E)*1000.
Analyzed by generalized estimating equations with identity link function and
exchangeable correlation structure.
Model 1 was adjusted for treatment and randomization stratum, model 2 was adjusted for
randomization stratum
43
Table 26. Association between change in plasma IFLs and change in BMI
Model 1 Model 2
Estimate (SE) P Estimate (SE) P
Daidzein 0.04 (0.07) 0.49 0.05 (0.07) 0.42
Dihydrodaidzein -0.10 (0.10) 0.46 -0.08 (0.11) 0.51
O-desmethylangolensin 0.09 (0.12) 0.43 0.10 (0.12) 0.37
Equol -0.14 (0.08) 0.25 -0.13 (0.08) 0.27
Genistein 0.03 (0.05) 0.51 0.04 (0.05) 0.43
O-glycitein -0.89 (0.76) 0.34 -0.84 (0.7) 0.37
Dihydrogenistein -0.44 (0.19) 0.09 -0.43 (0.19) 0.09
NOTE: Mean(E)*1000.
Analyzed by generalized estimating equations with identity link function and
exchangeable correlation structure.
Model 1 was adjusted for treatment and randomization stratum, model 2 was adjusted for
randomization stratum
Table 27. Association between change in plasma IFLs and change in LDL-cholesterol
Model 1 Model 2
Estimate (SE) P Estimate (SE) P
Daidzein -1.32 (1.27) 0.33 -1.66 (1.24) 0.21
Dihydrodaidzein -4.32 (2.21) 0.12 -4.08 (2.23) 0.08
O-desmethylangolensin 0.43 (1.86) 0.82 -0.30 (1.92) 0.88
Equol 3.46 (3.79) 0.22 2.98 (3.57) 0.29
Genistein -2.04 (1.25) 0.13 -2.27 (1.20) 0.08
O-glycitein -24.49 (20.24) 0.23 -26.81 (20.44) 0.19
Dihydrogenistein -4.01 (3.63) 0.38 -4.39 (3.59) 0.34
NOTE: Mean(E)*1000.
Analyzed by generalized estimating equations with identity link function and
exchangeable correlation structure.
Model 1 was adjusted for treatment and randomization stratum, model 2 was adjusted for
randomization stratum
44
Table 28. Association between change in plasma IFLs and change in triglycerides
Model 1 Model 2
Estimate (SE) P Estimate (SE) P
Daidzein 4.58 (2.89) 0.12 1.63 (2.75) 0.56
Dihydrodaidzein 6.04 (5.26) 0.27 2.11 (5.31) 0.70
O-desmethylangolensin 3.99 (4.11) 0.35 -0.68 (3.95) 0.86
Equol -3.37 (4.56) 0.48 -5.84 (5.03) 0.24
Genistein 5.42 (2.36) 0.04 2.59 (2.24) 0.28
O-glycitein 25.62 (43.36) 0.51 9.08 (37.71) 0.80
Dihydrogenistein 15.48 (9.19) 0.11 12.74 (8.92) 0.18
NOTE: Mean(E)*1000.
Analyzed by generalized estimating equations with identity link function and
exchangeable correlation structure.
Model 1 was adjusted for treatment and randomization stratum, model 2 was adjusted for
randomization stratum
Table 29. Association between change in plasma IFLs and change in glucose
Model 1 Model 2
Estimate (SE) P Estimate (SE) P
Daidzein -1.00 (1.01) 0.30 -0.15 (0.80) 0.85
Dihydrodaidzein -4.30 (4.23) 0.32 -2.14 (3.43) 0.55
O-desmethylangolensin 0.97 (2.22) 0.67 1.79 (2.30) 0.45
Equol 3.45 (3.14) 0.17 3.79 (3.34) 0.15
Genistein -1.19 (0.85) 0.15 -0.39 (0.70) 0.58
O-glycitein 9.50 (11.63) 0.41 13.91 (13.63) 0.27
Dihydrogenistein 2.98 (2.34) 0.23 4.25 (2.70) 0.11
NOTE: Mean(E)*1000.
Analyzed by generalized estimating equations with identity link function and
exchangeable correlation structure.
Model 1 was adjusted for treatment and randomization stratum, model 2 was adjusted for
randomization stratum
45
Table 30. Association between change in plasma IFLs and change in weight
Model 1 Model 2
Estimate (SE) P Estimate (SE) P
Daidzein 0.21 (0.37) 0.56 0.23 (0.36) 0.50
Dihydrodaidzein -0.58 (0.58) 0.43 -0.53 (0.59) 0.47
O-desmethylangolensin 0.49 (0.69) 0.45 0.54 (0.67) 0.40
Equol -0.83 (0.46) 0.25 -0.78 (0.46) 0.26
Genistein 0.15 (0.27) 0.56 0.18 (0.27) 0.49
O-glycitein -5.70 (4.16) 0.28 -5.47 (4.21) 0.30
Dihydrogenistein -2.54 (1.07) 0.09 -2.49 (1.08) 0.09
NTOE: Mean(E)*1000
Analyzed by generalized estimating equations with identity link function and
exchangeable correlation structure.
Model 1 was adjusted for treatment and randomization stratum, model 2 was adjusted for
randomization stratum
Table 31. Association between change in plasma IFLs and change in waist to hip ratio
Model 1 Model 2
Estimate (SE) P Estimate (SE) P
Daidzein 0.002 (0.002) 0.28 0.0019 (0.0019) 0.31
Dihydrodaidzein -0.0033(0.0036) 0.34 -0.0034 (0.0035) 0.32
O-desmethylangolensin -0.0031 (0.0034) 0.35 -0.0031 (0.0034) 0.35
Equol -0.0007 (0.003) 0.82 -0.0008 (0.003) 0.80
Genistein 0.0030 (0.0017) 0.07 0.0028 (0.0016) 0.08
O-glycitein 0.0143 (0.0292) 0.60 0.0139 (0.0292) 0.62
Dihydrogenistein 0.0001 (0.0039) 0.99 0.0000 (0.0038) 0.99
NOTE: Mean(E)*1000
Analyzed by generalized estimating equations with identity link function and
exchangeable correlation structure.
Model 1 was adjusted for treatment and randomization stratum, model 2 was adjusted for
randomization stratum
46
Table 32. Association between change in plasma IFLs and change in hip circumference
Model 1 Model 2
Estimate (SE) P Estimate (SE) p
Daidzein -0.08 (0.08) 0.32 -0.08 (0.08) 0.29
Dihydrodaidzein -0.09 (0.13) 0.54 -0.10 (0.13) 0.49
O-desmethylangolensin 0.14 (0.16) 0.40 0.12 (0.16) 0.45
Equol -0.16 (0.16) 0.29 -0.17 (0.17) 0.27
Genistein -0.16 (0.06) 0.01 -0.16 (0.06) 0.01
O-glycitein -0.78 (1.12) 0.46 -0.82 (1.12) 0.44
Dihydrogenistein -0.59 (0.24) 0.01 -0.60 (0.24) 0.01
NOTE: Mean(E)*1000.
Analyzed by generalized estimating equations with identity link function and
exchangeable correlation structure.
Model 1 was adjusted for treatment and randomization stratum, model 2 was adjusted for
randomization stratum
Table 33. Association between change in plasma IFLs and change in waist circumference
Model 1 Model 2
Estimate (SE) P Estimate (SE) p
Daidzein -0.01 (0.08) 0.90 0.002 (0.08) 0.98
Dihydrodaidzein 0.27 (0.18) 0.19 0.28 (0.18) 0.16
O-desmethylangolensin 0.05 (0.16) 0.77 0.07 (0.16) 0.68
Equol 0.22 (0.11) 0.15 0.23 (0.11) 0.12
Genistein 0.01 (0.07) 0.85 0.02 (0.07) 0.72
O-glycitein 0.41 (1.36) 0.77 0.48 (1.33) 0.73
Dihydrogenistein 0.57 (0.27) 0.01 0.58 (0.27) 0.01
NOTE: Mean(E)*1000.
Analyzed by generalized estimating equations with identity link function and
exchangeable correlation structure.
Model 1 was adjusted for treatment and randomization stratum, model 2 was adjusted for
randomization stratum
47
Table 34 summarizes the previous interventional studies of soy isoflavones on the effect of anthropometric and
metabolic measurements in humans.
Table 34. Previous interventional studies on the effect of isoflavones in humans
Population N Intervention Comparison Duration Measures Results
Japanese
perimenopausal
women
23 61.8mg soy
isoflavones
daily
placebo 4 weeks T-cholesterol, LDL-
cholesterol, HDL-
cholesterol, VLDL-
cholesterol, Triglycerides
Total serum
cholesterol and LDL
cholesterol were
decreased in the
isoflavone group
compared to the
placebo group,
Perimenopausal
women aged 45-
55 years.
51 1), 1 dose of
20g soy protein
supplement
daily
2), 2 doses of
10g soy protein
supplement
daily
20g complex
carbohydrate
supplement
daily
Crossover
trial with 6
weeks of
each
treatment
period
Total cholesterol, LDL
cholesterol, HDL-
cholesterol, Triglycerides,
weight
Total cholesterol and
LDL cholesterol
were reduced in both
treatment groups. No
effect on HDL
cholesterol or
triglycerides
African
American and
White
postmenopausal
women
216 20g soy protein
containing
160mg
isoflavones
daily
20g whole
milk protein
12 weeks Total cholesterol, LDL-
cholesterol, HDL-
cholesterol, LDL particle
number, Triglycerides
Total cholesterol,
LDL cholesterol and
LDL particle number
decreased in the
treatment group
compared to placebo
at 6
weeks
48
Table 34. Previous interventional studies on the effect of isoflavones in humans, continued
Hypercholestero
lemic
postmenopausal
women
66 1), 40g protein
containing
55.6mg of
isoflavones
daily
2), 40g protein
containing
90mg of
isoflavones
daily
casein and
nonfat dry
milk
6 months non-HDL cholesterol,
HDL-cholesterol, LDL
receptor mRNA
Non-HDL
cholesterol reduced,
HDL cholesterol
increased, and LDL
receptor mRNA
increased in both
ISP56 and ISP90
groups.
Postmenopausal
women
75 1), Soy protein
powder 30mg
containing
60mg of
isoflavones
daily
2), Soy
isoflavones
60mg (tablet)
daily
Casein
protein 30gm
daily
12 weeks Serum cholesterol, HDL
cholesterol, triglycerides,
LDL cholesterol, serum
apolipoprotein,
atherogenic index, serum
leutenizing hormone,
serum follicle-stimulating
hormone
Compared to the
placebo group,
HDL-cholesterol
level increased and
serum cholesterol,
triglycerides, LDL-
cholesterol levels
decreased in the soy
protein group;
triglycerides level
reduced in the soy
isoflavone extract
group.
Postmenopausal
women who had
been treated for
breast cancer
64 114mg of
isoflavones
daily
Similar
looking
placebo
tablets
Crossover
trial with 3
months of
each
treatment
period
Total cholesterol, LDL
cholesterol, HDL-
cholesterol, HDL-2
cholesterol, triglycerides,
apolipoprotein B,
Apolipoprotein A1,
lipoprotein
No effect was found
on lipids or
apolipoproteins
49
Table 34. Previous interventional studies on the effect of isoflavones in humans, continued
Healthy
postmenopausal
women aged 45
to 70
117 Cereal bar
containing
50mg
isoflavone
extract daily
Cereal bar
containing no
isoflavones
Crossover
trial with 8
weeks of
each
treatment
period
Total cholesterol, LDL
cholesterol, HDL-
cholesterol, total:HDL,
Triacylglycerol,
Lipoprotein, glucose,
percentage sdLDL,
nonesterified fatty acids
No effect was found
on lipids
Healthy
postmenopausal
women aged 60
to 75
202 99mg
isoflavones
daily
Milk protein 12 months Total cholesterol, HDL
cholesterol, LDL
cholesterol, triglycerides,
lipoprotein
No effect was found
Healthy obese
postmenopausal
women
100 80mg soy
isoflavone
extract daily
placebo 6 months anthropometric measures,
body composition, leptin,
adiponectin, TNF-alpha,
homocysteine, C-reactive
protein, glucose, insulin,
lipid profile and
oestradiol serum levels,
Kupperman index and
Cervantes Scale
No effect on lipids
or anthropometric
measures were found
Normocholester
olemic and
mildly
hypercholesterol
emic
postmenopausal
women
18 1), 65mg of
isoflavones
daily;
2), 132mg of
isoflavones
daily
7.1mg of
isoflavones
daily
Crossover
trial with 93
days of each
treatment
period
Total cholesterol, HDL
cholesterol, LDL
cholesterol,
triacylglycerol, total:HDL
cholesterol, LDL:HDL
cholesterol, apoprotein
A1, apoprotein b,
lipoprotein, LDL-PPD
132mg isoflavones
reduced plasma
LDL-cholesterol by
6.5% compared to
placebo, 65mg
isoflavones did not
significantly reduce
LDL-cholesterol
concentration;
Abstract (if available)
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
Association of bone density and bone metabolism markers: the Women’s Isoflavone Soy Health (WISH) trial
PDF
Associations between isoflavone soy protein (ISP) supplementation and breast cancer in postmenopausal women in the Women’s Isoflavone Soy Health (WISH) clinical trial
PDF
Soy isoflavone supplements for the treatment of menopausal hot flashes: the Women’s Isoflavone Soy Health (WISH) trial
PDF
Association between endogenous sex hormone levels and cognition: the Women’s Isoflavone Soy Health (WISH) trial
PDF
Physical activity and sex hormone levels in postmenopausal women
PDF
Effect of estradiol on circulating levels of inflammatory cytokines in postmenopausal women
PDF
Associations between physical activities with bone mineral density in postmenopausal women
PDF
The relationship of resistin, leptin, adiponectin, and ghrelin with bone mineral density in healthy postmenopausal women: longitudinal analysis
PDF
Sex hormones and atherosclerosis in postmenopausal women
PDF
Effects of post-menopausal hormone therapy on arterial stiffness in the ELITE trial
PDF
Associations between inflammatory markers and change in cognitive endpoints
PDF
Effect of menopausal hormone replacement therapy on lipoprotein particle fractions and association with carotid artery intima-media thickness and grey-scale median, independent measurements of su...
PDF
The associations of inflammatory markers with progression of subclinical atherosclerosis in early and late postmenopausal women
PDF
The effect of vitamin D supplementation on the progression of carotid intima-media thickness and arterial stiffness in elderly African American women: Results of a randomized placebo-controlled trial
PDF
Association of arterial stiffness progression with subclinical atherosclerosis measurements in postmenopausal women
PDF
Relationship between progression of atherosclerosis and coagulation measures in a randomized-controlled trial
PDF
Relationship of blood pressure and antihypertensive medications to cognitive change in the BVAIT, WISH, and ELITE clinical trials
PDF
The effectiveness of cycling intervention on cerebral palsy patients, and a survey validation method
PDF
The association between self-reported physical activity and cognition in elderly clinical trial participants
PDF
Association of carotid artery stiffness with measures of cognition in older adults
Asset Metadata
Creator
Wang, Jun
(author)
Core Title
Effect of soy isoflavones on anthropometric and metabolic measurements in postmenopausal women
School
Keck School of Medicine
Degree
Master of Science
Degree Program
Applied Biostatistics
Publication Date
01/24/2011
Defense Date
12/12/2010
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
anthropometric,clinical trials,isoflavones,lipid,metabolic,OAI-PMH Harvest,postmenopausal women,soy
Place Name
USA
(countries)
Language
English
Contributor
Electronically uploaded by the author
(provenance)
Advisor
Mack, Wendy J. (
committee chair
), Azen, Stanley Paul (
committee member
), Hodis, Howard Neil (
committee member
)
Creator Email
jwang1206@gmail.com,wangjun@usc.edu
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-m3621
Unique identifier
UC1279629
Identifier
etd-Wang-4238 (filename),usctheses-m40 (legacy collection record id),usctheses-c127-427727 (legacy record id),usctheses-m3621 (legacy record id)
Legacy Identifier
etd-Wang-4238.pdf
Dmrecord
427727
Document Type
Thesis
Rights
Wang, Jun
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
Tags
anthropometric
clinical trials
isoflavones
lipid
metabolic
postmenopausal women
soy