The Journal of Nutrition
Soy Protein Intake Has Sex-Specific Effects on
the Risk of Metabolic Syndrome in Middle-Aged
and Elderly Chinese1–3
An Pan,4Oscar H. Franco,5Jianping Ye,6Wendy Demark-Wahnefried,7Xingwang Ye,4Zhijie Yu,4
Huaixing Li,4and Xu Lin4*
4Key Laboratory of Nutrition and Metabolism, Institute for Nutritional Sciences, Shanghai Institutes for Biological Sciences, Chinese
Academy of Sciences and Graduate School of the Chinese Academy of Sciences, Shanghai, 200031, China;5Senior Public Health
Epidemiologist, Unilever Corporate Research, Colworth House, Sharnbrook, Bedfordshire, MK441LQ, United Kingdom;
6Pennington Biomedical Research Center, Louisiana State University, Baton Rouge, LA 70808; and7University of Texas-MD Anderson
Cancer Center, Houston, TX 77030
Soy protein intake has been postulated to improve lipid profiles, glucose homeostasis, and blood pressure. However, data
linking soy protein intake and metabolic syndrome (MetS) are limited. We evaluated the association between soy protein
intake and the risk of MetS and its components among middle-aged and elderly Chinese. A cross-sectional study was
conducted among 2811 Chinese men and women aged 50–70 y, who were free of diagnosed cardiovascular diseases and
cancers. Dietary data, including soy protein intake, was collected using a 74-item FFQ. MetS was defined using the
updated National Cholesterol Education Program Adult Treatment Panel III criteria for Asian-Americans. We used
multivariate logistical regression models to quantify these associations. The median level of soy protein intake was 7.82 g/d
(7.64 g/d inmenand 8.02g/d inwomen).Overall,the associationofsoy protein intakeand the riskofMetSdifferedbetween
men and women (P for interaction ¼ 0.008). In men, the adjusted odds ratio comparing the extreme quartiles was 1.64
(95% CI: 0.95–2.81; P-trend ¼ 0.077), whereas for women, it was 0.66 (95% CI: 0.42–1.03; P-trend ¼ 0.138). Soy protein
intake was positively associated with hyperglycemia (P-trend ¼ 0.005) in men, whereas it was inversely associated with
elevated blood pressure (P-trend ¼ 0.049). It was not associated with any component in women. In conclusion, habitual
soy protein intake may have sex-dependent effects on risk of MetS in middle-aged and elderly Chinese.J. Nutr. 138:
Metabolic syndrome (MetS)8is a cluster of metabolic abnor-
malities of which central obesity, hyperglycemia, hypertriglyc-
eridemia, decreased HDL cholesterol, and elevated blood pressure
are major characters (1). Existing data suggest that the incidence
of the MetS is rising at an alarming rate globally (2). Using the
updated National Cholesterol Education Program Adult Treat-
ment Panel III criteria for Asian-Americans (3), we reported
previously that the prevalence of MetS was ;40% in middle-
aged and elderly Chinese (4). Several other studies also found
that a large proportion of Chinese adults suffered from MetS,
which has become an important public health concern in China
Genetic, metabolic, and environmental factors, including
diet, are thought to play major roles in the etiology of MetS (7).
Soy foods containing soy protein and isoflavones (i.e. daidzein
and genistein, major soy-based phytoestrogens) have gained consid-
erable attention for their potential effects on improving cardio-
vascular risk factors, including lipid profiles (8), glucose and
insulin homeostasis (9), weight control (10), and blood pressure
(11). Therefore, it is hypothesized that the intake of soy foods
and soy protein may be associated with a decreased risk of MetS.
This proposition has been tested in animal studies (12–14) and
1 clinical trial (15). Azadbakht et al. (15) found that a 30-g/d soy
regimen of soy protein or soy nuts as a replacement for red meat
in the Dietary Approaches to Stop Hypertension diet signifi-
cantly improved glycemic control and lipid profiles in 42
postmenopausal women with MetS.
intake of 10 g/d (16,17). Several cross-sectional epidemiological
1Supported by the Knowledge Innovation Program Project of the Chinese
Academy of Sciences (KSCX1-YW-02, KSCX2-2-25, and KSCX2-YW-R-116), the
and the Shanghai-Unilever Research Development Fund (CH-2006-0941).
2Supplemental Tables 1–3 and Supplemental Figure 1 are available with the
online posting of this paper at jn.nutrition.org.
3Author disclosures: A. Pan, O. Franco, J. Ye, W. Demark-Wahnefried, X. Ye, Z.
Yu, H. Li, and X. Lin, no conflicts of interest.
8Abbreviations used: IQR, interquartile range; MetS, metabolic syndrome; OR,
odds ratio; SHBG, sex hormone-binding globulin.
* To whom correspondence should be addressed. E-mail: email@example.com.
0022-3166/08 $8.00 ª 2008 American Society for Nutrition.
Manuscript received 5 August 2008. Initial review completed 3 September 2008. Revision accepted 8 September 2008.
by guest on October 13, 2011
Supplemental Material can be found at:
studies in Chinese and Japanese populations found that greater
miological reports are available on the association of habitual
consumption of soy foods or soy protein with MetS in the general
population. Therefore, we took advantage of the comprehensive
data collected in the Nutrition and Health of Aging Population in
China project to investigate the relationship between soy protein
and elderly Chinese.
Subjects and Methods
Study participants. The present study is part of the Nutrition and
Health of Aging Population in China project, which was a population-
based cohort study among noninstitutionalized Chinese people aged
50–70 y who were stable residents for at least 20 y in Beijing (north) or
Shanghai (south), China. Inbrief,thisstudywasconductedsimultaneously
in both locations from March to June 2005; 1 rural site and 2 urban
sites were selected in each location to obtain an adequate sampling of
individuals from lower, middle, and upper socioeconomic status. Only
1 participant per household was randomly selected and individuals were
excluded from the study if they had 1 of the following conditions: self-
care disabilities; severe psychological disorders; tuberculosis or other
communicable diseases; AIDS; or had a diagnosis of cancer, coronary
heart disease, stroke, Alzheimer’s disease, or dementia within the 6-mo
period before the start of the study. Further details of the study are
published elsewhere (4). The study was approved by the Institutional
Review Board of the Institute for Nutritional Sciences and informed
consent was obtained from each participant.
From a residency registration list in the randomly selected streets or
villages, 3533 potential participants were identified and successfully
interviewed in their households and 3379 persons agreed to participate;
complete data (i.e. questionnaires, physical examinations, and blood
samples) were obtained for 3289 participants. For the current analyses,
we excluded participants who had missing or incomplete dietary data or
those with implausibly high (.16.7 MJ) or low (,2.5 MJ) total daily
energy intakes (n ¼ 124). In addition, we excluded those with self-
reported coronary heart disease, stroke, or cancer (n ¼ 354). After these
exclusions, 2811 subjects remained for the final analyses.
General questionnaire interview. In-house, personal interviews were
conducted by trained research staff using a standardized questionnaire
(4). Sociodemographic variables included age, gender, geographic
location (Beijing/Shanghai), residential region (urban/rural), marital
status (yes/no), employment status (employed/retired/ unemployed, or
status (yes/no) and alcohol drinking status (yes/no) also were assessed. The
physical activity level for each individual was classified as low, moderate,
or high according to the International Physical Activity Questionnaire
(short last 7 d format) scoring protocol with minor modification (22).
Participants also were asked whether they were diagnosed with dia-
betes, dyslipidemia, or hypertension and whether they had taken hypo-
lipidemic, antihypertensive,ororalantidiabeticagents, and/orinsulin.For
the purposes of this study, we defined comorbidity as having any 1 of the
advanced diseases or taking any of the aforementioned medications.
Dietary assessment. Data on nutrient intake in the year prior to
enrollment were derived from a quantitative FFQ that was administered
during a personal interview. The FFQ included 74 food items/groups
(Supplemental Table 1) and was modified from the 2002 China Health
and Nutrition Survey (23). Seven soy food items responsible for the
fresh and dried soybeans, tofu, soy milk, jelly bean curd, soy sauce, and
other processed soy products. Study participants were asked to recall the
frequency of consumption of individual food items (number of times per
day, week, month, and year) and estimated portion size, using local weight
units [liang (50 g)] or natural units (cup, bowl) per unit time. The Chinese
Food Composition Table (24) was used to estimate the intake levels of
major nutrients for study participants. Total dietary intake of each nutrient
was calculated by adding the intakes of that nutrient from the information
collected for each food item. Nutrient intake for an individual food item
Composition Table with the reported frequency and the amount of intake.
residual method (25) and the energy-adjusted values are presented.
MetS definition. All participants were invited to have a physical
examination at the local health stations or the community clinics after
the home interview. Participants were required to fast overnight ($7 h)
before the physical examination by trained staff and physicians using
standard protocols. Measurements of body weight and height, waist and
hip circumference, and blood pressure have been described previously
(4). BMI levels were categorized as normal weight (,24.0 kg/m2),
overweight ($24.0 kg/m2), or obese ($28.0 kg/m2) according to the
criteria established for Chinese populations (26).
Peripheral venous EDTA-plasma samples were collected and centri-
fuged at 2500 3 g; 15 min at 4?C. After being frozen, the samples were
shipped in dry ice to the Institute for Nutritional Sciences and stored at
280?Cuntil analysis. Plasma total cholesterol, HDLand LDL cholesterol,
triglycerides, and glucose were measured enzymatically on Hitachi
automatic analyzer with reagents purchased from Wako Pure Chemical
MetS was defined using the updated National Cholesterol Education
Program Adult Treatment Panel III criteria for Asian-Americans (3), i.e.
presenting $3 of the following conditions: 1) central obesity: waist
circumference $90 cm in men or $80 cm in women; 2) hypertriglyc-
eridemia: triglycerides $1.7 mmol/L; 3) low HDL cholesterol: HDL
cholesterol , 1.03 mmol/L in men or , 1.30 mmol/L in women;
4) elevated blood pressure: blood pressure $130/85 mm Hg or current
use of antihypertensive medications; 5) hyperglycemia: fasting plasma
glucose $5.6 mmol/L, or previously diagnosed type 2 diabetes or on oral
antidiabetic agents or insulin.
Statistical analysis. Study participants were divided into 4 categories
by quartile distribution of total soy protein intake. The lowest quartile
served as the reference group. Because of the skewed distribution of soy
protein intake, we reported the median and interquartile range (IQR) per
quartile. Distributions of sociodemographic information, anthropomet-
ric measures, dietary and lifestyle characteristics, and MetS risk factors
were presented for each quartile.
In multivariate logistical regressions, we evaluated the association
between total soy protein intake and MetS, adjusting for relevant and
important demographic and biological confounders (not collinear) and
other covariates suggested by previous publications (18–21).
All models included age, geographic location, residential region,
current drinking and smoking status, marital status, educational level,
employment status, BMI, physical activity level, and presence of
comorbidity. Dietary factors included in the final models were total
energy intake, dietary fat, cholesterol, nonsoy protein, and fiber intake.
We also conducted analyses stratified by geographic location, resi-
dential region, and gender. When analyzing the association in women,
menopausal status (yes/no) was also included in the final model.
Analyses also were stratified by the presence or absence of comorbidity
to determine whether diagnosis of diabetes, dyslipidemia, or hyperten-
sion would modify the associations. Tests of linear trend across in-
creasing quartiles of intake were conducted using the median value of
each quartile and treating it as a continuous variable. All statistical tests
were based on 2-sided probability using Stata 9.2 (StataCorp).
Distributions of demographic characteristics and dietary
information linked to MetS. Participants with MetS were
2414Pan et al.
by guest on October 13, 2011
older (58.8 6 5.8 y) than those without MetS (58.2 6 6.1 y; P ¼
0.015) (Table 1). The prevalence of Met S was also significantly
higher in female participants and those who were from Beijing,
those living in urban communities, those who were overweight or
obese, and those having comorbidities. No significant associations
were observed with marital status, alcohol drinking, educational
attainment, physical activity level, total energy intake, total fat,
Selected characteristics of study participants according to
the quartiles of total soy protein intake. The median level
of the soy protein intake was 7.82 g/d (IQR ¼ 4.65–12.60 g/d,
7.64 g/d in men and 8.02 g/d in women) (Table 2). Higher intakes
ofsoy protein werereported byurbanresidents.Participants with
high soy protein intake also tended to have a higher level of
education and household income, a higher BMI, lower levels of
physical activity, and a higher prevalence of comorbidity.
Greater consumption of soy protein was associated with higher
intakes of dietary fat, protein, fiber, and cholesterol, and a lower
intake of carbohydrate. Soy protein consumption was not
associated with total energy intake.
Association between soy protein intake and MetS. Overall,
there was no significant association between soy protein intake
and the risk of MetS after adjustment for potential confounding
factors (Table 3). The association between soy protein intake
and the risk of MetS, however, differed markedly between men
and women (P for interaction ¼ 0.008). Among male partici-
pants, the intake of soy protein was moderately associated with
increased risk of MetS after adjusting for potential demographic
and biological covariates (P-trend ¼ 0.077) (Table 3; Supple-
mental Fig. 1A). The adjusted odds ratio (OR) between the
highest and lowest quartiles of soy protein intake was 1.64 (95%
CI ¼ 0.95–2.81). However, the adjusted OR between extreme
quartiles was 0.66 (95% CI ¼ 0.42–1.03) in women. We also
evaluated the risk of MetS using sex-specific quartiles of soy pro-
tein intake and the results did not differ (Supplemental Table 2).
There were no interactions between soy protein intake and
geographic location (north/south), residential region (urban/
rural), and presence of comorbidity (yes/no). In these subgroups,
Association between soy protein intake and MetS com-
ponents. In the overall study population, soy protein intakewas
associated with increased risk of hyperglycemia (P-trend ¼
0.007) and marginally reduced risk of elevated blood pressure
(P-trend ¼ 0.056) (Table 4; Supplemental Fig. 1B,C). However,
soy protein intake was not associated with other components
(central obesity, hypertriglyceridemia, and low HDL choles-
terol) (Table 4; Supplemental Fig. 1D–F).
There were significant interactions between soy protein
intake and sex in the association with hypertriglyceridemia
(P for interaction ¼ 0.005), low HDL cholesterol level (P for
interaction ¼ 0.060), and hyperglycemia (P for interaction ¼
0.081). The positive association between soy protein intake and
hyperglycemia was primarily confined to males (P-trend ¼
0.005). Men in the top quartile of soy protein intake had an89%
increased risk of hyperglycemia compared with the lowest quar-
tile (OR ¼ 1.89; 95% CI ¼ 1.22–2.92). Excluding those with
diagnosed diabetes or those taking antidiabetic treatment did
not substantially change the results (data not shown). Soy protein
intake and elevated blood pressure were inversely associated in
men (P-trend ¼ 0.049). Soy protein intake was not significantly
associated with any component of MetS in women. We also
calculated the association between sex-specific quartiles of soy
protein intake and MetS components and the results were ba-
sically similar (Supplemental Table 3).
In this cross-sectional study of middle-aged and elderly Chinese,
a higher intake of soy protein tended to be associated with a
reduced risk of MetS in women but an elevated risk in men. To
our knowledge, this is the first population-based study assessing
the correlation between regular soy protein intake and risk of
MetS, a strong predictor of cardiovascular diseases and type 2
diabetes. The findings of substantial sex-dependent effects of soy
protein on MetS, particularly glucose and lipid homeostasis, are
novel and warrant further studies.
Soybeans and soy products are functional foods with a
complex milieu of potentially active substances, such as soy
protein, isoflavones, fiber, and polyunsaturated fats. Obviously,
it was not possible to exclude the effects of other soy compo-
nents in the present study; therefore, we also evaluated the effect
of the total soy amount and isoflavones [according to the score
method advocated by de Kleijn et al. (27) or using the database
developed by Thompson et al. (28)] on MetS. The results were
consistent with soy protein: sex-specific effects of soy amount
and isoflavones on MetS and its components, particularly on
glucose homeostasis and lipid profile (data not shown). How-
ever, it still should be borne in mind that the statistical analysis
might not be completely accurate without a standard chemical
analysis of isoflavone content in food, given the marked vari-
ability of isoflavones in various foods.
Although most previous studies have focused on only
individual features of metabolic disorder such as dyslipidemia
or hypertension, several investigators also observed variations in
magnitude and/or direction related to intake of soy protein or
isoflavonesbetweenmen and women (19,20,29,30). For instance,
Zhang et al. (20) suggested that the association between soy
intake and bifurcation intima-media thickness differed between
Nagata et al. (29) found an inverse association of soy product
intake with diastolic blood pressure in Japanese men and a
marginally positive association in peri- and postmenopausal
women. Indeed, these sex-specific effects of soy have also been
observed in male and female animals (31,32). Stauffer et al. (32)
demonstrated that soy isoflavone supplementation augmented
cardiac growth in the hypertrophic cardiomyopathic male mice
and attenuated growth in the females.
of soy may be linked to its estrogen-like activity. Soy isoflavones,
as a major category of phytoestrogen, can bind to the estrogen
receptors a and b, with a higher affinity to estrogen receptor b
(33). Existing evidence has shown that isoflavones could act as
estrogen agonists or antagonists, depending on the target tissues
(34), doses (35) and types (36) of isoflavones, and endogenous
circulating sex hormone profile (37). Indeed, a previous cross-
sectional study in Shanghai reported that soy intake was asso-
ciated differently with glycosuria between premenopausal and
postmenopausal women, indicating that menopausal status and
endogenous estrogen levels may modify the phytoestrogen
activity of soy (21). In addition, the sex-specific effects of sexual
hormones on glucose and lipid metabolism and MetS have been
demonstrated previously (38). High testosterone concentrations
are shown to be associated with a lower risk of type 2 diabetes in
men but with a higher risk in women; whereas sex hormone-
binding globulin (SHBG), which is inversely associated with risk
Soy protein intake and metabolic syndrome 2415
by guest on October 13, 2011
of diabetes, has a stronger effect on women than on men (39). A
Isoflavones may directly modulate concentrations of circulating
sex hormones and SHBG in a sex-specific fashion. In fact, epi-
demiological evidence suggests an inverse association between
serum endogenous androgen concentrations and soy consump-
Japanese dietary pattern, rich in soybean products and vegeta-
bles, was positively associated with impaired glucose tolerance
(43), whereas phytoestrogen intake was found to be positively
associated with SHBG in a cross-sectional study in postmeno-
pausal women (44), and several clinical trials observed increased
SHBG after soy protein or isoflavone supplements (45,46). Con-
sequently, soy protein with isoflavones may have adverse effects
in males while having favorable effects in females. However,
further large epidemiological studies and long-term clinical trials
are correlated on regulating glucose and lipid metabolism.
ior by gender or obesity status when completing dietary question-
naires (25). However, studies in Chinese men and women found
that both men and women overestimate their soy intake, which
suggests ‘‘a bias of study participants seeking social approval’’
(47,48). Such systematic bias should not substantially alter our
observed effects of soy protein on MetS and its components,
because such bias would just attenuate the associations (25).
Other potential mechanisms might involve nonestrogenic
effects of isoflavones that could interfere with glucose metab-
olism in the males. Genistein, as a major kind of isoflavones, has
General characteristics of the 2811 participants
according to the presence or absence of MetS
and OR (95%CI) for MetS1
(n ¼ 1146)
(n ¼ 1665) OR (95% CI)
50–54 (n ¼ 909)
55–59 (n ¼ 713)
60–64 (n ¼ 605)
65–70 (n ¼ 584)
Male (n ¼ 1173)
Female (n ¼ 1638)
Beijing (n ¼ 1305)
Shanghai (n ¼ 1506)
Urban (n ¼ 1378)
Rural (n ¼ 1433)
Yes (n ¼ 2463)
No (n ¼ 348)
Educational level, y in school
#9 (n ¼ 2167)
.9 (n ¼ 644)
Employed (n ¼ 691)
Retired (n ¼ 1521)
Unemployed or on
welfare (n ¼ 599)
No (n ¼ 2048)
Yes (n ¼ 763)
Current alcohol drinker
No (n ¼ 2050)
Yes (n ¼ 761)
Physical activity level
Low (n ¼ 211)
Moderate (n ¼ 1176)
High (n ¼ 1424)
Presence of comorbidity
No (n ¼ 1789)
Yes (n ¼ 1022)
(n ¼ 1348)
Overweight or obesity
(n ¼ 1463)
Total energy intake
Q1 (#7.43 MJ/d)
Q2 (7.44–9.05 MJ/d)
Q3 (9.06–11.1 MJ/d)
Q4 ($11.2 MJ/d)
Total fat intake
Q1 (#56.5 g/d)
Q2 (56.6–68.6 g/d)
218 (19.0)1130 (67.9)1.00 (ref.)6
928 (81.0)535 (32.1)8.62 (7.16–10.37)
(n ¼ 1146)
(n ¼ 1665)
OR (95% CI)
Q3 (68.7–82.2 g/d)
Q4 ($82.3 g/d)
Q1 (#201 mg/d)
Q2 (202–315 mg/d)
Q3 (316–474 mg/d)
Q4 ($475 mg/d)
Nonsoy protein intake
Q1 (#53.6 g/d)
Q2 (53.7–60.4 g/d)
Q3 (60.5–67.7 g/d)
Q4 ($67.8 g/d)
Dietary carbohydrate intake
Q1 (#303.7 g/d)
Q2 (303.8–336.1 g/d)
Q3 (336.2–371.4 g/d)
Q4 ($371.5 g/d)
Dietary fiber intake
Q1 (#9.68 g/d)
Q2 (9.69–12.4 g/d)
Q3 (12.5–15.3 g/d)
Q4 ($15.4 g/d)
1Values are n (%).
2Adjusted for gender, geographic location, and residential region.
3Adjusted for age (continuous), geographic location, and residential region.
4Adjusted for age (continuous), gender, and residential region.
5Adjusted for age (continuous), gender, and geographic location.
6Adjusted for age (continuous), gender, geographic location, and residential region.
7All nutrients are energy-adjusted using residual regression method.
2416 Pan et al.
by guest on October 13, 2011
been shown to act as a potent and direct inhibitor of glucose
transporter 4-mediated glucose uptake (49). Genistein was also
found to specifically inhibit tyrosine kinase (50), suggesting it
could inhibit the tyrosine kinase activity of insulin receptor (51).
Moreover, previous animal studies have shown that genistein
lowered the threshold of glucose-induced insulin secretion in
islets and rapidly increased the blood insulin (52). In addition,
genistein was found to affect adipose tissue deposition in a dose-
dependent andsex-specific mannerandinduced peripheralinsulin
resistance in male mice (53), whereas it produced antiadipogenic
effects in female ovariectomized mice (54). A recent study also
increased peripheral insulin resistance and decreased plasma
adiponectin in 91 adult male cynomolgus monkeys (55). How-
ever, it still remains unknown whether similar dose- and sex-
dependent effects could manifest in humans, and the long-term
exposure of habitual soy protein and isoflavones on glucose and
lipid homeostasis in humans still needs further elucidation.
In the present study, we observed that soy protein intake was
associated with a reduced risk of hypertension in men. This
finding is consistent with a cross-sectional study of a Japanese
population in which a mild association of soy intake with blood
pressure reduction was found in men but not in postmenopausal
women (29). A clinical trial conducted by Jenkins et al. (30) also
Characteristics of the 2811 subjects by quartiles of soy protein intake1
Quartile of soy protein intake
Variable Total Q1 Q2Q3Q4
Median (IQR), g/d
Educational level (years in school)
Current alcohol drinker
High physical activity level
Presence of comorbidity
Systolic BP, mm Hg
Diastolic BP, mm Hg
Plasma total cholesterol, mmol/L
Plasma LDL cholesterol, mmol/L
Plasma HDL cholesterol, mmol/L
Plasma triglycerides, mmol/L
Fasting plasma glucose, mmol/L
Total energy intake, MJ/d
Dietary fat intake,2g/d
Dietary cholesterol intake,2mg/d
Nonsoy protein intake,2g/d
Dietary fiber intake,2g/d
2811702703 703 703
7.82 (4.65–12.60) 2.97 (1.68–3.85)6.19 (5.46–6.98)9.86 (8.72–11.07)16.34 (14.16–20.95)
58.4 6 6.0
57.7 6 5.7
59.1 6 6.1
58.6 6 6.0
58.3 6 6.0
24.4 6 3.6
83.3 6 10.6
139.0 6 22.1
79.7 6 10.7
4.69 6 0.98
3.25 6 0.97
1.28 6 0.33
1.36 6 1.05
5.82 6 1.75
9.42 6 2.66
69.8 6 21.0
348.1 6 216.1
61.4 6 12.3
336.0 6 57.5
13.0 6 5.3
24.2 6 3.6
83.3 6 10.7
141.4 6 22.0
81.0 6 10.7
4.58 6 0.96
3.12 6 0.94
1.29 6 0.33
1.33 6 1.11
5.72 6 1.62
10.46 6 2.77
61.9 6 23.8
267.2 6 230.9
56.7 6 11.4
361.7 6 65.2
10.3 6 4.3
24.2 6 3.7
82.2 6 10.9
140.3 6 22.0
79.7 6 10.4
4.65 6 0.93
3.22 6 0.94
1.30 6 0.32
1.25 6 0.85
5.76 6 1.61
8.82 6 2.57
68.6 6 19.0
332.7 6 191.3
60.5 6 11.6
340.7 6 51.3
12.3 6 4.4
24.3 6 3.4
83.1 6 10.2
138.4 6 22.2
79.3 6 10.9
4.73 6 0.97
3.31 6 0.99
1.28 6 0.34
1.38 6 1.01
5.91 6 1.97
8.87 6 2.34
71.5 6 17.9
375.2 6 194.3
63.2 6 10.9
330.3 6 49.5
13.3 6 4.2
24.8 6 3.7
84.4 6 10.4
136.0 6 21.9
78.9 6 10.9
4.79 6 1.04
3.36 6 0.99
1.26 6 0.34
1.50 6 1.19
5.89 6 1.75
9.53 6 2.59
77.0 6 19.9
417.2 6 216.8
65.2 6 13.5
311.3 6 50.7
16.3 6 6.2
1Unless otherwise specified, data are expressed as n (%) or mean 6 SD.
2All nutrients are energy-adjusted using residual regression method.
Soy protein intake and metabolic syndrome2417
by guest on October 13, 2011
indicated thatdiets high insoysignificantlyreduced systolicblood
pressure in men but not in women. However, several studies
found either minor or no associations between soy protein or
isoflavone intake and blood pressure in either men or women
[summarized in (29)]. Thus, this blood pressure-lowering effect
still needs further research.
Numerous studies have investigated the effects of soy protein
and isoflavones on cardiometabolic risk factors and have yielded
inconsistent results. A meta-analyses of randomized clinical trials
indicated that high-dose isolated or concentrated soy protein
supplementation generally had a favorable effect on serum lipid
the diet improvedfeatures ofthe MetSinpostmenopausalwomen
(15). Nonetheless, the AHA Nutrition Committee released a
scientific advisory on soy protein and cardiovascular disease in
2006, suggesting that ‘‘a very large amount of soy protein, more
few percentage points (? 3%). No benefit is evident on HDL
cholesterol, triglycerides, lipoprotein(a) or blood pressure. Thus,
supplements is minimal at best (56).’’ Compelling evidence sug-
gests that dietary composition modification has substantial effects
losing weight and improving lipid profiles and insulin resistance
term adherence to these diets was generally poor (59,60). Indeed,
there is still much controversy inthe area ofnutrition intervention
and its components (type of protein, fat, and carbohydrate) may
influence the course of chronic disease and warrant further
we suggest that further studies consider gender differences.
The major strength of this study is that the sample was
population based and heterogeneous. The samples represent pop-
China. Thus, our findings should be generalizable. Another
strength is that all of the subjects consumed significant amounts
of soy protein in their daily life with a considerable variation in
intake, thus facilitating the detection of the potential effects.
Furthermore, the soy and other exposure assessments were
conducted ;1 wk before the physical examination and blood
sample collection; thus, the exposure evaluations were unlikely to
be influenced by health-related outcomes.
The primary limitation of this study is that the FFQ used was
not specifically validated. Nevertheless, the FFQ was modified
from the validated 2002 China Health andNutrition Survey (23)
and the 7 soy food items included within the FFQ are similar to
those used in several other studies in Chinese populations
(16,19,20). Furthermore, the median level of soy protein intake
was comparable to those studies, suggesting that our FFQ was
indeed reliable. Another limitation of the present study is the
cross-sectional study design, in which the temporal sequence of
soy food intake and development of MetS cannot be fully
inferred. Moreover, sex hormone concentrations were not mea-
sured in our participants; therefore, we could not test whether sex
hormonesmediate the association between soy proteinintakeand
MetS. Finally, although we have controlled for several dietary
and other potential confounding factors suggested to be asso-
ciated with MetS and soyfood intake, we cannot fully exclude
the possibility that other residual confounders may affect the
results. For instance, it remains unclear to what extent the rapid
nutritional transition and westernization of the Chinese culture
could modify these associations.
In conclusion, we found that habitual soy protein intake
conferred sex-specific effects on risk of MetS, with moderately
reduced risk in middle-aged and elderly women but elevated risk
in men. It significantly increased risk for hyperglycemia but
reduced risk for hypertension in middle-aged and elderly men.
The study may represent the first epidemiological study of the
effect of soy products on MetS in humans. Additional work is
needed to confirm these results and determine the underlying
OR and 95% CI for MetS associated with soy protein intake
Quartile of soy protein intakeP–value for
Q1Q2 Q3 Q4
Total (n ¼ 2811)
Men (n ¼ 1173)
Women (n ¼ 1638)
1P-value for linear trend was calculated by assigning median value to each quartile and treating as continuous variable.
2Adjusted for geographic location (north/south), residential region (urban/rural), sex, age, educational level (#9 or .9 y in school), marital
status (yes/no), employment status (employed, retired, unemployed/on welfare), current smoking and alcohol-use (yes/no), physical activity
level (low/moderate/high), presence of comorbidity (yes/no). For men and women, sex was not adjusted. For women, menopausal status
(yes/no) was included in the final model.
3Further adjusted for total energy intake, dietary fat, fiber, nonsoy protein, and cholesterol intake (all in quartiles).
4Further adjusted for BMI.
5P for interaction ¼ 0.006 (model 1), 0.005 (model 2), and 0.008 (model 3).
2418Pan et al.
by guest on October 13, 2011
mechanisms for the differential impact of soy protein on men
We are grateful to Dr. Eric Ding for his valuable comments on
manuscript preparation and revision.
1. Eckel RH, Grundy SM, Zimmet PZ. The metabolic syndrome. Lancet.
Kereiakes DJ, Willerson JT. Metabolic syndrome epidemic. Circulation.
Grundy SM, Cleeman JI, Daniels SR, Donato KA, Eckel RH, Franklin
BA,Gordon DJ, Krauss RM, Savage PJ, et al. Diagnosis and manage-
ment of the metabolic syndrome: an American Heart Association/
National Heart, Lung, and Blood Institute Scientific Statement. Circu-
Ye X, Yu Z, Li H, Franco OH, Liu Y, Lin X. Distributions of C-reactive
protein and its association with metabolic syndrome in middle-aged and
older Chinese people. J Am Coll Cardiol. 2007;49:1798–805.
Gu D, Reynolds K, Wu X, Chen J, Duan X, Reynolds RF, Whelton PK,
He J, InterASIA Collaborative Group. Prevalence of the metabolic
syndrome and overweight among adults in China. Lancet. 2005;365:
He Y, Jiang B, Wang J, Feng K, Chang Q, Fan L, Li X, Hu FB.
Prevalence of the metabolic syndrome and its relation to cardiovascular
disease in an elderly Chinese population. J Am Coll Cardiol. 2006;47:
7. Poulsen P, Vaag A, Kyvik K, Beck-Nielsen H. Genetic versus environ-
mental aetiology of the metabolic syndrome among male and female
twins. Diabetologia. 2001;44:537–43.
Zhan S, Ho SC. Meta-analysis of the effects of soy protein containing
isoflavones on the lipid profile. Am J Clin Nutr. 2005;81:397–408.
Bhathena SJ, Velasquez MT. Beneficial role of dietary phytoestrogens in
obesity and diabetes. Am J Clin Nutr. 2002;76:1191–201.
10. Cope MB, Erdman JW Jr, Allison DB. The potential role of soyfoods in
weight and adiposity reduction: an evidence-based review. Obes Rev.
11. Appel LJ. The effects of protein intake on blood pressure and cardio-
vascular disease. Curr Opin Lipidol. 2003;14:55–9.
12. Dyrskog SE, Jeppesen PB, Colombo M, Abudula R, Hermansen K.
Preventive effects of a soy-based diet supplemented with stevioside on
the development of the metabolic syndrome and type 2 diabetes in
Zucker diabetic fatty rats. Metabolism. 2005;54:1181–8.
13. Davis J, Iqbal MJ, Steinle J, Oitker J, Higginbotham DA, Peterson RG,
Banz WJ. Soy protein influences the development of the metabolic
syndrome in male obese ZDFxSHHF rats. Horm Metab Res. 2005;37:
14. Davis J, Higginbotham A, O’Connor T, Moustaid-Moussa N, Tebbe A,
Kim YC, Cho KW, Shay N, Adler S, et al. Soy protein and isoflavones
influence adiposity and development of metabolic syndrome in the
obese male ZDF rat. Ann Nutr Metab. 2007;51:42–52.
15. Azadbakht L, Kimiagar M, Mehrabi Y, Esmaillzadeh A, Padyab M, Hu
FB, Willett WC. Soy inclusion in the diet improves features of the
metabolic syndrome: a randomized crossover study in postmenopausal
women. Am J Clin Nutr. 2007;85:735–41.
16. Chen Z, Zheng W, Custer LJ, Dai Q, Shu XO, Jin F, Franke AA. Usual
dietary consumption of soy foods and its correlation with the excretion
OR and 95% CI for each component of MetS associated with soy protein intake1
Quartile of soy protein intakeP–value for
Total (n ¼ 2811)3
Elevated blood pressure
Low HDL cholesterol
Men (n ¼ 1173)3
Elevated blood pressure
Low HDL cholesterol6
Women (n ¼ 1638)3
Elevated blood pressure
Low HDL cholesterol6
1Hyperglycemia was defined as a fasting plasma glucose $5.6 mmol/L or previously diagnosed type 2 diabetes or taking oral antidiabetic
agents or insulin. Elevated blood pressure was defined as blood pressure $130/85 mm Hg, or current use of antihypertensive medications.
Central obesity was defined as waist circumference $90 cm in men or $80 cm in women. Hypertriglyceridemia was defined as
triglycerides $1.7 mmol/L. Low HDL cholesterol was defined as HDL cholesterol , 1.03 mmol/L in men or , 1.30 mmol/ L in women.
2P-value for linear trend was calculated by assigning median values to each quartile and treating as a continuous variable.
3Adjusted for geographic location (north/south), residential region (urban/rural), sex, age, educational level (#9 or .9 y in school), marital
status (yes/no), employment status (employed, retired, unemployed/on welfare), current smoking and alcohol use (yes/no), physical activity
level (low/moderate/high), body mass index, presence of comorbidity (yes/no), and dietary information (total energy intake, dietary fat, fiber,
nonsoy protein, and cholesterol intake, all in quartiles). For men and women, sex was not adjusted. For women, menopausal status (yes/no)
was included in the final model.
4P for interaction ¼ 0.081.
5P for interaction ¼ 0.005.
6P for interaction ¼ 0.060.
Soy protein intake and metabolic syndrome2419
by guest on October 13, 2011
rate of isoflavonoids in overnight urine samples among Chinese women
in Shanghai. Nutr Cancer. 1999;33:82–7.
17. Nagata C, Takatsuka N, Shimizu H. Soy and fish oil intake and
mortality in a Japanese community. Am J Epidemiol. 2002;156:
18. Nagata C, Takatsuka N, Kurisu Y, Shimizu H. Decreased serum total
cholesterol concentration is associated with high intake of soy products
in Japanese men and women. J Nutr. 1998;128:209–13.
19. Ho SC, Woo JL, Leung SS, Sham AL, Lam TH, Janus ED. Intake of soy
products is associated with better plasma lipid profiles in the Hong
Kong Chinese population. J Nutr. 2000;130:2590–3.
20. Zhang B, Chen YM, Huang LL, Zhou XX, Chen CG, Ye YB, Su YX.
Greater habitual soyfood consumption is associated with decreased
carotid intima-media thickness and better plasma lipids in Chinese
middle-aged adults. Atherosclerosis. 2008;198:403–11.
21. Yang G, Shu XO, Jin F, Elasy T, Li HL, Li Q, Huang F, Zhang XL,
Gao YT, et al. Soyfood consumption and risk of glycosuria: a cross-
sectional study within the Shanghai Women’s Health Study. Eur J Clin
22. Craig CL, Marshall AL, Sjostrom M, Bauman AE, Booth ML,
Ainsworth BE, Pratt M, Ekelund U, Yngve A, et al. International
physical activity questionnaire: 12-country reliability and validity. Med
Sci Sports Exerc. 2003;35:1381–95.
23. Zhao W, Hasegawa K, Chen J. The use of food-frequency question-
naires for various purposes in China. Public Health Nutr. 2002;5:
24. Yang Y. Chinese Food Composition Table 2002. Beijing: Peking Univer-
sity Medical Press; 2003.
25. Willett WC. Nutritional epidemiology. New York: Oxford University
26. Zhou BF, Cooperative Meta-Analysis Group of the Working Group on
Obesity in China. Predictive values of body mass index and waist
circumference for risk factors of certain related diseases in Chinese
adults: study on optimal cut-off points of body mass index and waist
circumference in Chinese adults. Biomed Environ Sci. 2002;15:83–96.
27. de Kleijn MJ, van der Schouw YT, Wilson PW, Adlercreutz H, Mazur W,
Grobbee DE, Jacques PF. Intake of dietary phytoestrogens is low in
postmenopausal women in the United States: the Framingham study.
J Nutr. 2001;131:1826–32.
28. Thompson LU, Boucher BA, Liu Z, Cotterchio M, Kreiger N.
Phytoestrogen content of foods consumed in Canada, including
isoflavones, lignans, and coumestan. Nutr Cancer. 2006;54:184–201.
29. Nagata C, Shimizu H, Takami R, Hayashi M, Takeda N, Yasuda K.
Association of blood pressure with intake of soy products and other
food groups in Japanese men and women. Prev Med. 2003;36:692–7.
30. Jenkins DJ, Kendall CW, Jackson CJ, Connelly PW, Parker T, Faulkner
D, Vidgen E, Cunnane SC, Leiter LA, et al. Effects of high- and low-
isoflavone soyfoods on blood lipids, oxidized LDL, homocysteine, and
blood pressure in hyperlipidemic men and women. Am J Clin Nutr.
31. Blair RM, Appt SE, Bennetau-Pelissero C, Clarkson TB, Anthony MS,
Lamothe V, Potter SM. Dietary soy and soy isoflavones have gender-
specific effects on plasma lipids and isoflavones in golden Syrian f(1)b
hybrid hamsters. J Nutr. 2002;132:3585–91.
32. Stauffer BL, Konhilas JP, Luczak ED, Leinwand LA. Soy diet worsens
heart disease in mice. J Clin Invest. 2006;116:209–16.
33. Morito K, Hirose T, Kinjo J, Hirakawa T, Okawa M, Nohara T,
Ogawa S, Inoue S, Muramatsu M, et al. Interaction of phytoestrogens
with estrogen receptors alpha and beta. Biol Pharm Bull. 2001;24:
34. Safe SH, Pallaroni L, Yoon K, Gaido K, Ross S, Saville B, McDonnelle
D. Toxicology of environmental estrogens. Reprod Fertil Dev. 2001;13:
35. Hsieh CY, Santell RC, Haslam SZ, Helferich WG. Estrogenic effects of
genistein on the growth of estrogen receptor-positive human breast
cancer (MCF-7) cells in vitro and in vivo. Cancer Res. 1998;58:3833–8.
36. Mueller SO, Simon S, Chae K, Metzler M, Korach KS. Phytoestrogens
and their human metabolites show distinct agonistic and antagonistic
properties on estrogen receptor alpha (ERalpha) and ERbeta in human
cells. Toxicol Sci. 2004;80:14–25.
37. Glazier MG, Bowman MA. A review of the evidence for the use of
phytoestrogens as a replacement for traditional estrogen replacement
therapy. Arch Intern Med. 2001;161:1161–72.
38. Regitz-Zagrosek V, Lehmkuhl E, Mahmoodzadeh S. Gender aspects of
the role of the metabolic syndrome as a risk factor for cardiovascular
disease. Gend Med. 2007;4 Suppl B:S162–77.
39. Ding EL, Song Y, Malik VS, Liu S. Sex differences of endogenous sex
hormones and risk of type 2 diabetes: a systematic review and meta-
analysis. JAMA. 2006;295:1288–99.
40. Sutton-Tyrrell K, Wildman RP, Matthews KA, Chae C, Lasley BL,
Brockwell S, Pasternak RC, Lloyd-Jones D, Sowers MF, et al. Sex-
hormone-binding globulin and the free androgen index are related to
cardiovascular risk factors in multiethnic premenopausal and perimen-
opausal women enrolled in the Study of Women Across the Nation
(SWAN). Circulation. 2005;111:1242–9.
41. Nagata C, Inaba S, Kawakami N, Kakizoe T, Shimizu H. Inverse
association of soy product intake with serum androgen and estrogen
concentrations in Japanese men. Nutr Cancer. 2000;36:14–8.
42. Low YL, Taylor JI, Grace PB, Dowsett M, Folkerd E, Doody D,
Dunning AM, Scollen S, Mulligan AA, et al. Polymorphisms in the
CYP19 gene may affect the positive correlations between serum and
urine phytoestrogen metabolites and plasma androgen concentrations in
men. J Nutr. 2005;135:2680–6.
43. Mizoue T, Yamaji T, Tabata S, Yamaguchi K, Ogawa S, Mineshita M,
Kono S. Dietary patterns and glucose tolerance abnormalities in
Japanese men. J Nutr. 2006;136:1352–8.
44. Low YL, Dunning AM, Dowsett M, Folkerd E, Doody D, Taylor J,
Bhaniani A, Luben R, Khaw KT, et al. Phytoestrogen exposure is
associated with circulating sex hormone levels in postmenopausal
women and interact with ESR1 and NR1I2 gene variants. Cancer
Epidemiol Biomarkers Prev. 2007;16:1009–16.
J Jr, Sepkovic DW, BradlowHL, et al.Effect of soy protein on endogenous
hormones in postmenopausal women. Am J Clin Nutr. 2002;75:145–53.
46. Pino AM, Valladares LE, Palma MA, Mancilla AM, Yanez M, Albala C.
Dietary isoflavones affect sex hormone-binding globulin levels in
postmenopausal women. J Clin Endocrinol Metab. 2000;85:2797–800.
47. Villegas R, Yang G, Liu D, Xiang YB, Cai H, Zheng W, Shu XO.
Validity and reproducibility of the food-frequency questionnaire used in
the Shanghai Men’s Health Study. Br J Nutr. 2007;97:993–1000.
48. Shu XO, Yang G, Jin F, Liu D, Kushi L, Wen W, Gao YT, Zheng W.
Validity and reproducibility of the food frequency questionnaire used in
the Shanghai Women’s Health Study. Eur J Clin Nutr. 2004;58:17–23.
49. Bazuine M, van den Broek PJ, Maassen JA. Genistein directly inhibits
GLUT4-mediated glucose uptake in 3T3–L1 adipocytes. Biochem
Biophys Res Commun. 2005;326:511–4.
50. Akiyama T, Ishida J, Nakagawa S, Ogawara H, Watanabe S, Itoh N,
Shibuya M, Fukami Y. Genistein, a specific inhibitor of tyrosine-specific
protein kinases. J Biol Chem. 1987;262:5592–5.
51. Joost HG. Structural and functional heterogeneity of insulin receptors.
Cell Signal. 1995;7:85–91.
52. Sorenson RL, Brelje TC, Roth C. Effect of tyrosine kinase inhibitors on
islets of Langerhans: evidence for tyrosine kinases in the regulation of
insulin secretion. Endocrinology. 1994;134:1975–8.
53. Penza M, Montani C, Romani A, Vignolini P, Pampaloni B, Tanini A,
Brandi ML, Alonso-Magdalena P, Nadal A, et al. Genistein affects
adipose tissue deposition in a dose-dependent and gender-specific
manner. Endocrinology. 2006;147:5740–51.
54. Naaz A, Yellayi S, Zakroczymski MA, Bunick D, Doerge DR, Lubahn
DB, Helferich WG, Cooke PS. The soy isoflavone genistein decreases
adipose deposition in mice. Endocrinology. 2003;144:3315–20.
55. Wagner JD, Zhang L, Shadoan MK, Kavanagn K, Chen H, Tresnasari
K, Kaplan JR, Adams MR. Effects of soy protein and isoflavones on
insulin resistance and adiponectin in male monkeys. Metabolism. 2008;
57 Suppl 1:S24–31.
56. Sacks FM, Lichtenstein A, Van Horn L, Harris W, Kris-Etherton P,
Winston M. Soy protein, isoflavones, and cardiovascular health: an
American Heart Association Science Advisory for professionals from
the Nutrition Committee. Circulation. 2006;113:1034–44.
57. McAuley KA, Hopkins CM, Smith KJ, McLay RT, Williams SM, Taylor
RW, Mann JI. Comparison of high-fat and high-protein diets with a
high-carbohydrate diet in insulin-resistant obese women. Diabetologia.
58. Gardner CD, Kiazand A, Alhassan S, Kim S, Stafford RS, Balise RR,
Kraemer HC, King AC. Comparison of the Atkins, Zone, Ornish, and
LEARN diets for change in weight and related risk factors among
2420 Pan et al.
by guest on October 13, 2011
overweight premenopausal women: the A TO Z Weight Loss Study: a Download full-text
randomized trial. JAMA. 2007;297:969–77.
59. Brinkworth GD, Noakes M, Keogh JB, Luscombe ND, Wittert GA,
Clifton PM. Long-term effects of a high-protein, low-carbohydrate diet
on weight control and cardiovascular risk markers in obese hyper-
insulinemic subjects. Int J Obes Relat Metab Disord. 2004;28:661–70.
60. Dansinger ML, Gleason JA, Griffith JL, Selker HP, Schaefer EJ.
Comparison of the Atkins, Ornish, Weight Watchers, and Zone diets
for weight loss and heart disease risk reduction: a randomized trial.
61. Feldeisen SE, Tucker KL. Nutritional strategies in the prevention and
Soy protein intake and metabolic syndrome2421
by guest on October 13, 2011