ORIGINAL ARTICLE

A mild favorable effect of soy protein with isoflavoneson body compositionFa 6-month double-blindrandomized placebo-controlled trial among Chinesepostmenopausal women

Z-m Liu1,2, SC Ho1,2, Y-m Chen1,2,3 and YP Ho4

1School of Public Health and Primary Care, The Chinese University of Hong Kong, Hong Kong SAR; 2Center of Research andPromotion of Women’s Health, School of Public Health and Primary Care, The Chinese University of Hong Kong, Hong KongSAR; 3Department of Medical Statistical and Epidemiology, School of Public Health, Sun Yat-sen University, Guangzhou,China and 4School of Pharmacy, The Chinese University of Hong Kong, Hong Kong SAR

Background: In vitro and animal studies suggested that soy protein and isoflavones promote weight and body fat loss. However,clinical trials in humans were few and the effects remained uncertain.Objectives: To examine whether soy protein with isoflavones and isoflavone extracts exert beneficial effects on bodycomposition among postmenopausal women.Design: A double-blind randomized placebo-controlled trial was conducted among 180 postmenopausal Chinese women withmild hyperglycemia. After a 2-week adaptation, participants were randomly assigned to one of the three arms to receive 15 gsoy protein and 100 mg isoflavones (Soy group), or 15 g milk protein and 100 mg isoflavones (Iso group), or 15 g milk protein(placebo group) on a daily basis for 6 months.Results: We observed a mild but significant favorable effect of soy protein with isoflavones on the changes of body weight(BW), body mass index, and body fat percentage relative to isoflavone extracts and milk protein after 6-month supplementation.The mean differences of BW between the Soy and the Iso groups were �0.75 kg (95% CI: �1.363 to �0.136, P¼0.017), andbetween the Soy and the placebo groups were �0.60 kg (95% CI: �1.209 to �0.019, P¼0.047). The mean difference ofchange percentage in body fat percentage between the Soy and the Iso groups was �3.74% (95% CI: �6.88 to �0.60%,P¼0.02), and between the Soy and the placebo groups was �2.54% (95% CI: �5.69 to 0.12%. P¼ 0.08).Conclusions: Six-month supplementation of soy protein with isoflavones had a mild favorable effect on body composition inpostmenopausal women.

International Journal of Obesity advance online publication, 17 November 2009; doi:10.1038/ijo.2009.236

Keywords: soy protein; isoflavones; body weight; body fat percentage; body mass index

Introduction

Aging and menopause-induced estrogen deficiency is related

to several changes that contribute to an increased risk of

cardiovascular disease. These include an increase in body

weight (BW) and fat mass (FM) and changes in fat

metabolism.1,2 Studies on postmenopausal hormone therapy

(HT) have indicated the beneficial effects on body composi-

tion, such as preventing the menopausal-related abdominal

fat increase, with preservation of muscle strength.3,4 As HT is

associated with some adverse health effects,5,6 there is

increasing interest to find an alternative to HT to prevent

and delay the menopause-related health risks.

Soy is a traditional Asian diet. Soybean contains many

beneficial components, among which soy protein and

isoflavones have received most research attention. Current

evidence from in vitro and animal studies suggest that soy

protein and isoflavones may promote weight loss and

prevent body fat gain, with additional benefits over otherReceived 13 July 2009; revised 21 September 2009; accepted 5 October 2009

Correspondence: Professor SC Ho, Centre of Research and Promotion of

Women’s Health, School of Public Health and Primary Care, Faculty of

Medicine, The Chinese University of Hong Kong, 4/F School of Public Health

and Primary Care, Prince of Wales Hospital, Shatin, N.T 852, Hong Kong.

E-mail: suzanneho@cuhk.edu.hk

International Journal of Obesity (2009) 1–10& 2009 Macmillan Publishers Limited All rights reserved 0307-0565/09 $32.00

www.nature.com/ijo

protein sources.7–9 Epidemiologic observations also indi-

cated that persons with high soy food consumption have

lower BWs than those with low soy intakes.10,11 A cross-

sectional study in western postmenopausal women10 sugges-

ted a significant inverse relation between isoflavone

consumption and BW, body mass index (BMI), waist

circumference (WC), and total body FM (assessed by bio-

electric impedance analysis). However, thus far, there have

been only limited human trials investigating the long-term

effects of dietary soy protein on body composition and

weight loss.

Although some randomized clinical trials (RCTs) suggested

a beneficial effect of soy consumption on body composi-

tion,12–14 inconsistent findings have been noted in other

studies.15,16 The discrepancies among studies may be attribu-

ted to the variations in population characteristics, types of

soy products, treatment duration, or inadequate sample size

and random errors. As most studies were designed to compare

the effects of soy protein with or without isoflavones,12,15,16

the independent effects of isoflavones and its combined

effect with soy protein were not delineated within the same

study. Thus, the hypothesis of a beneficial effect of soy intake

on body composition is still speculative.

This paper reports the effects of soy protein with iso-

flavones and isoflavone extracts on changes in body

composition among postmenopausal Chinese women with

early untreated hyperglycemia. We hypothesized that the

effect from soy protein together with isoflavones was more

pronounced than isoflavone extracts alone.

Patients and methods

Participant recruitment

Potential participants were recruited from the local commu-

nity through multiple strategies, including advertisements in

local newspapers, health talk, written invitations, flyers, or

subjects referrals. A total of 180 postmenopausal women

with prediabetes or early diabetes were recruited to partici-

pate in a 6-month RCT. Subjects were Hong Kong Chinese

menopausal women aged 48–70 years without a menstrual

cycle for at least 1 year. Potential participants were excluded

if they had a history of breast, uterine, or ovarian cancer, or

stroke, coronary heart disease, or thyroid disease in the

previous 5 years. They were also excluded if they currently,

or in preceding 3 months, were on medication treatment for

weight, glucose, lipids reduction, HT, or had a known allergy

to soy or milk. The Ethical Committee of the Chinese

university of Hong Kong approved the study and all the

participants gave written informed consent before their

enrollment.

Study design and intervention regimes

The study was a 6-month double-blind randomized placebo-

controlled trial. Before formal admission to the study,

subjects received 2 weeks of placebo, 15 g milk protein

(Pacific Dairy Ingredients Co., Ltd, Shanghai), as run-in to

get familiarized with the study requirement and procedure.

Subjects with good compliance and without any side effects

were randomly assigned to one of the three groups:

Soy protein (Solpro 931) and isoflavones (Solgen 40/s) were

provided by Solbar Industries Ltd, Ashdod, Israel. Baseline

measurements were performed before randomization. The

intake compliance was assessed by counting the empty and

remaining sachets at each face-to-face interview and estimat-

ing the percentage of supplements provided that had been

consumed.

Randomization and blinding

Block randomization in block intervals of 15 was used for

subject randomization. A list of random numbers was

computer generated. Investigators not involved in subject

contact, data collection, and analysis performed the randomi-

zation and the labeling work. The research staff assigned the

number series to the subject in the order of their enrollment

into the trial. Participants, investigators, and laboratory

technicians were blinded to the treatment assignment until

the conclusion of the trial. To assess the efficacy of blinding,

the participants were asked, at the end of the intervention,

to which group they thought they had been assigned. The

proportion of the participants thought they were assigned to

either one of the treatments was similar among the three

treatment groups, indicating a successful blinding efficacy

(data not shown).

Supplements preparation and delivery

The three kinds of preparation were formulated into iso-

caloric powder with similar color, taste, and odor as well as

similar nutrients profile (protein, fat, carbohydrate, magne-

sium, calcium, sodium, etc.) by the addition of vitamins,

minerals, and other additives. The daily dose was filled into

identical looking unmarked packet. Each sachet contained

25 g powdered supplements, of which 60% was protein. A

measure of 100 mg isoflavones, expressed in aglycone form,

contained 35 mg daidzin, 59 mg genistin, and 4 mg glycitin,

with 98.5% in conjugated form. Supplements were suggested

to be mixed with 300 ml of water or any other liquid of

subject’s choice and detailed instruction on how to prepare

the supplements was given by oral and video demonstration.

To avoid taste fatigue, recipes were provided on how to

include the supplement into usual diet. Subjects were asked

to consume one packet daily, preferably at breakfast.

Supplements were delivered to subjects at intervals of 1, 2,

and 3 months after randomization.

Placebo group 15 g milk protein

Iso group 15 g milk protein+100 mg isoflavones

Soy group 15 g soy protein+100 mg isoflavones

Soy protein and body compositionZ-m Liu et al

2

International Journal of Obesity

Power of study

With a planned number of subjects of 60 in each arm (180

subjects in total), we would have 80% power to find an

improvement of 1.3% in the change of BW, and 3% in the

change of body fat percentage (BF%), based on a conven-

tional assumption of a level of 0.05 (for a two-sided t-test)

and allowing for 10% withdrawal rate.

Data collection

Data were collected by trained interviewers through face-to-

face interview based on a structured and previously validated

questionnaire on sociodemographic data, years since meno-

pause (YSM), medical history, medication treatment, dietary

habits, smoking, alcohol drinking, and physical activities.

Dietary intake and physical activity assessment

During the course of intervention, dietary intake was

measured by 3-day food records that were completed by

subjects before each follow-up visit at 0, 3, and 6 months.

Subjects received a 30-min training about utensil sizes, food

amounts, and portion estimation. A set of visual aids

including a food photo album and common household

containers were also provided to help them estimate the

amounts of the foods and beverage taken. Dietary nutrient

intakes were calculated based on the China Food Composi-

tion Table (2004).17 Subjects were asked not to take

supplements containing phytoestrogens or other extracts

known to affect outcome measures. Participants’ vitamin

and mineral supplement use were recorded at the baseline

visit and they were instructed to make no change in the

doses during the study period. Participants were instructed

to maintain their habitual diet and level of physical activity

for the duration of the study.

Anthropometric measurements

BW was measured to the nearest 0.1 kg, with the subjects in

light clothing without shoes with the use of beam balance

scale (Detecto, Clinicon Medical Ltd, Ontario, CA, USA).

Height was measured in a standing position without shoes

using a wall-mounted stadio-meter. BMI was calculated as

weight in kilograms divided by height in meters squared.

WC was measured at the level of midway between the lower

rib margin and the iliac crest using a non-elastic tape

without any pressure to the body surface. Hip circumference

was measured at the level of the greater trochanters. WHR

was calculated as the ratio of waist-to-hip circumferences. All

measurements were performed twice and the average was

recorded.

BF%, FM, and free FM were determined by bioelectrical

impedance analyzer (BIA, TBF-410-GS Tanita Body Composi-

tion Analyzer, Tokyo, Japan). The BF% was calculated from

the impedance value as well as the pre-entered personal

particulars, including weight, height, age, and sex through

the built-in software.18 These measurements were done after

10–12 h fasting. Subjects were requested to refrain from

strenuous exercise 10–12 h before the measurements and

empty their bladder before the measurements. The co-

efficients of variation for repeated measures were all o5%

for BF%, FM, and free FM.

Statistical analysis

Statistical analysis was performed with the use of SPSS

(version 13.0). All results were considered significant if the

two-tailed P-value was o0.05. Baseline characteristics,

dietary intake, and physical activity level were compared

among the three study groups by analysis of variance

(ANOVA). Data were analyzed according to an intention-

to-treat principle including all 180 subjects. The last value

carried forward was used for any missing data at follow-ups.

Secondary analyses comprised a per protocol analysis, which

included only 164 subjects with good compliance (defined as

consuming X80% of provided supplements and done all

assessments and sample collections). The non-compliant

subjects were compared with the compliant subjects in

baseline characteristics by w2 or t-test.

A repeated-measures ANOVA was used to examine the

interactive effects of treatment and time. The major

approach to analysis was made by ANOVA and analysis of

covariance (general linear model) to compare the percentage

change at 3- and 6-month follow-ups among the three study

groups. Percentage changes in outcome measures were

calculated as ((follow-up value�baseline value)�100%/base-

line value). Least significant difference (LSD) test was used

for post hoc multiple comparisons. We also conducted

stratification analysis to examine whether the effect of soy

supplementation differed by YSM and baseline BMI levels.

Results

From November 2007 to April 2008, a total of 180 eligible

subjects who had successfully completed the 2-week run-in

period were randomized into the three treatment arms

(Figure 1). Subjects who withdrew from the study were also

invited for follow-up visits. In all, 176 subjects attended the

mid-term visit and 175 attended the 6-month visit.

Dropout and compliance

In all, 18 women withdrew during the 2-week run-in because

of loss of interests or time constraint, and 14 women (7.8%)

withdrew during the study after randomization. The most

important reasons for withdrawal after randomization were

gastrointestinal tract discomfort (n¼5), too busy to return

for follow-up assessment (n¼2), or developing some medical

conditions (eye disease or heavy sore throat) (n¼2). Detailed

reasons for the withdrawal were indicated in Figure 1.

Relatively more dropouts were observed in the placebo

Soy protein and body compositionZ-m Liu et al

3

International Journal of Obesity

group (n¼9) than the Soy (n¼3) and Iso (n¼2) groups.

Adherence was good and did not differ among the three

groups. More than 90% of the subjects used 97% of the

supplements.

Baseline characteristics

Baseline characteristics indicated that women in the three

groups were comparable in terms of age, YSM, job status,

education, family and medical history, smoking, dietary

nutrient intake, and physical activity pattern (Table 1).

Women lost to follow-up (n¼14) had similar baseline

characteristics with women retained in the study (data not

shown). Subjects’ dietary intake and physical activity

remained similar at baseline and at follow-ups (Table 1).

Effects of soy on body composition

At baseline, the mean values of the anthropometric measures

were similar among the three groups (Table 2). For the

repeated-measures ANOVA, we found marginal significance

in BW (P¼0.097), BMI (P¼0.101), and WHR (P¼0.104) in

the interaction terms between period and treatment. The

mean BW at 6 months was significantly lower in the Soy

group than the other two groups after adjusting for the

potential confounders (YSM, baseline BW, baseline energy

and soy protein intake, baseline physical activity level,

P¼0.042). A mild but significant decrease was found in

BW change percentage from baseline to 6 months in the Soy

group, as compared with the other two groups by either

ANOVA (P¼0.048) or analysis of covariance (P¼0.019).

Follow-up post hoc comparison indicated that the adjusted

mean difference at 6 months between the Soy and the Iso

groups was �0.75 kg (95% CI: �1.363 to �0.136, P¼0.017),

and between the Soy and the placebo groups was �0.60 kg

(95% CI: �1.209 to �0.019, P¼0.045). The adjusted mean

difference in percentage change of BW at 6 months was

�1.34% (95% CI: �2.344 to �0.333, P¼0.009) between the

Soy and the Iso groups, and �0.90% (95% CI: �1.911 to

0.011, P¼0.052) between the Soy and the placebo groups.

Similar tendency was observed in BMI, WC, WHR, BF%,

and FM, whereas statistical significance was found in BMI

(P¼0.048) and BF% (P¼0.042) at 6-month percentage

change after controlling for confounders. A marginal

significance was observed at 3-month change percentage in

BMI and BF%. For BF%, the adjusted mean difference in

change percentage between the Soy and the Iso groups was

�3.74% (95% CI: �6.88 to �0.60, P¼0.02), and between the

Soy and the placebo groups was �2.54% (95% CI: �5.69 to

0.12%. P¼0.08). No significant difference between the Iso

18 give up or excluded during 2-week run-in

Eligible subjects with pre or early DM: n=198

Randomized: n=180

Dropouts: n=3Reasons:Weight loss:1Skin allergy:1Menstrualrecurrence:1

Dropouts: n=2Reasons:Too busy:1Develop disease: 1Poor compliance: n=1

Dropouts: n=9Reasons:GI discomfort:5Too busy:1Feel hot:1Develop disease: 1Taking glucose-lowering drug: 1Poor compliance: n=1

Soy protein+Iso:n=60 Milk protein+Iso: n=60 Milk protein: n=60

After 6 m intervention

Valid completer: n=164

Included in finalanalysis: n=60

Included in finalanalysis: n=60

Included in finalanalysis: n=60

Completed : n=57 Completed : n=57 Completed : n=50

Figure 1 Participants flow diagram.

Soy protein and body compositionZ-m Liu et al

4

International Journal of Obesity

and the placebo groups was observed for all the anthropo-

metric indices in either 3- or 6-month change or percentage

change.

Further exploratory subgroup analysis using general linear

model controlling for possible covariates revealed that soy

protein together with isoflavones was more effective in

lowering BW (P¼0.023), BMI (P¼0.025), WC (P¼0.053),

and body FM (P¼0.053) in those within 4 years of their

menopause (Table 3).

Adverse events

We documented 75 reported adverse events, 25 in Soy group,

21 in Iso group, and 29 in placebo group. In all, 49.3%

(n¼35) complained of gastrointestinal discomfort, 25.3%

(n¼19) weight gain, and 14.7% (n¼11) felt ‘heatiness’ (dry

mouth, sore throat, acne, etc.). One person in the placebo

group reported pain from swollen breast and three (two in

Soy group, one in Iso group) suffered mild vaginal bleeding.

The number and types of adverse events did not differ

among the three groups.

Discussion

Our 6-month double-blind randomized placebo-controlled

trial demonstrated that, relative to isoflavone extracts and

milk protein, soy protein with isoflavones had a mild but

significant effect on the improvement of BW and body

composition in postmenopausal women, with responses

being more pronounced among those within 4 years of their

menopause.

Table 1 Baseline characteristics of 180 study participants by three treatment armsa

Soy group (n¼60) Iso group (n¼60) Placebo group (n¼60) P-value

Mean s.d. Mean s.d. Mean s.d.

Age (years) 56.4 4.7 56.0 4.4 55.9 3.8 0.816

YSM (years) 6.0 5.1 6.2 5.4 5.7 4.5 0.858

Age at menarche (years) 12.9 1.5 13.3 1.8 13.1 1.9 0.389

Job status 0.500

Housewife 28 46.7% 32 53.3% 34 56.7%

Part time 10 16.7% 9 15.0% 13 21.7%

Full time 21 35.0% 19 31.7% 13 21.7%

Education 0.774

Primary 11 6.1% 16 8.9% 15 8.3%

Middle school 41 22.8% 35 19.4% 36 20.0%

University 7 3.9% 9 5.0% 8 4.4%

Current use of hypotensive drugs 16 26.7% 14 23.3% 13 21.7% 0.634

Ever use of contraceptive drugs 32 53.3% 40 66.7% 32 53.3% 0.255

Ever use of HRT 13 21.6% 12 20.0% 11 18.3% 0.881

Passive smoking 12 20.0% 13 21.7% 16 26.7% 0.686

Physical activity (PA) level (mets-min per day)

Total PA at baseline 1216.6 513.7 1107.0 590.5 1109.3 527.9 0.460

Total PA at follow-ups on average 1125.5 456.1 1089.0 514.9 1096.9 488.3 0.385

Dietary nutrients intake at baseline

Total energy (kcal per day) 2202.5 778.7 2247.7 652.7 2025.3 762.1 0.217

Protein (g per day) 95.55 53.88 96.87 32.19 87.18 29.59 0.358

Total fat (g per day) 69.81 31.16 74.14 32.39 64.93 53.25 0.457

% kcal from fat 28.33 7.42 29.20 7.52 27.38 9.04 0.466

Soy protein (g per day) 12.31 16.80 12.40 9.66 12.55 11.81 0.996

Isoflavones (mg per day) 15.65 27.36 15.31 12.14 15.60 13.69 0.996

Dietary intake at follow-ups in average

Total energy (kcal per day) 2211.0 681.9 2222.5 517.0 1976.3 418.6 0.112

Protein (g per day) 95.3 35.7 96.9 25.0 85.3 17.7 0.059

Total fat (g per day) 71.5 28.4 71.6 22.6 67.1 19.3 0.311

% kcal from fat 28.5 5.1 28.7 5.0 27.4 6.1 0.413

Soy protein (g per day) 12.5 11.4 11.7 6.1 12.4 6.8 0.631

Isoflavones (mg per day) 16.8 15.2 16.4 7.8 17.2 8.0 0.491

Abbreviations: HRT, hormone replacement treatment; YSM, years since menopause. aData are presented as mean (s.d.) for continuous variable, or number (%) for

categorical variable. ANOVA for continuous variables and w2 test for categorical variables.

Soy protein and body compositionZ-m Liu et al

5

International Journal of Obesity

Table 2 Anthropometric indices at baseline and their change percentage at follow-ups in 180 intent-to-treat women

ITT Soy group (n¼60) Iso group (n¼ 60) Placebo group (n¼60) P* P**

Mean s.d. Mean s.d. Mean s.d.

Body weight (kg): P(T�G)¼0.097

Baseline 58.57 10.19 60.22 10.28 60.02 9.30 0.611

Month 3 58.15 10.34 60.04 10.23 60.37 10.07 0.439 0.126

Month 6 57.90 10.22 60.20 10.30 59.92 9.77 0.397 0.042

Change % (0–3 months) �0.78 2.58 �0.28 2.37 0.55 6.30 0.211 0.099

Change % (0–6 months) �1.18 3.02a,b,c,d �0.03 2.48 �0.25 2.57 0.048 0.019

Body mass index (kg m�2): P(T�G)¼ 0.101

Baseline 24.09 3.78 24.75 3.82 24.57 3.41 0.602

Month 3 23.92 3.83 24.68 3.83 24.70 3.70 0.436 0.652

Month 6 23.82 3.79 24.74 3.83 24.52 3.55 0.367 0.623

Change % (0–3 months) �0.78 2.58 �0.28 2.37 0.55 6.30 0.211 0.099

Change % (0–6 months) �1.00 2.74b,d �0.04 2.48 �0.28 2.56 0.109 0.048

Waist circumference (cm): P(T�G)¼0.127

Baseline 82.96 9.04 83.84 9.80 83.13 8.49 0.853

Month 3 82.41 9.46 83.42 9.81 83.01 8.53 0.838 0.658

Month 6 81.52 9.27 83.39 9.77 82.66 8.55 0.534 0.883

Change % (0–3 months) �0.68 3.72 �0.43 4.13 �0.10 2.98 0.680 0.809

Change % (0–6 months) �1.44 3.75 �0.47 3.73 �0.61 3.54 0.294 0.386

Hip circumference (cm): P(T�G)¼ 0.252

Baseline 94.17 7.32 96.17 8.71 96.11 6.73 0.266

Month 3 94.37 8.00 96.49 8.83 95.54 7.16 0.351 0.623

Month 6 94.09 7.71 96.47 8.94 95.59 7.61 0.268 0.417

Change % (0–3 months) 0.18 2.55 0.34 2.19 �0.58 3.07 0.130 0.192

Change % (0–6 months) �0.09 2.45 0.32 2.52 �0.55 3.30 0.233 0.232

Waist-to-hip ratio (WHR): P(T�G)¼ 0.104

Baseline 0.88 0.06 0.87 0.06 0.86 0.06 0.351

Month 3 0.87 0.06 0.86 0.05 0.87 0.05 0.609 0.385

Month 6 0.87 0.06 0.86 0.05 0.86 0.05 0.965 0.676

Change % (0–3 months) �0.83 3.85 �0.73 4.33 0.58 4.73 0.137 0.299

Change % (0–6 months) �1.40 3.72 �0.68 4.45 0.01 4.36 0.185 0.189

Body fat %: P(T�G)¼ 0.197

Baseline 31.58 6.63 32.96 8.14 31.92 6.05 0.533

Month 3 31.45 6.63 33.61 8.71 32.46 6.37 0.272 0.488

Month 6 31.02 6.94 33.46 8.38 32.22 6.63 0.193 0.296

Change % (0–3 months) �0.40 7.30 1.87 5.52 1.83 6.69 0.097 0.056

Change % (0–6 months) �1.75 9.60a,b,c,d 1.64 8.25 0.95 6.84 0.064 0.042

Fat mass (kg): P(T�G)¼0.271

Baseline 18.92 7.04 20.38 8.63 19.52 6.33 0.554

Month 3 18.76 6.88 20.86 9.09 19.80 6.62 0.322 0.266

Month 6 18.64 7.19 20.66 8.68 19.70 6.77 0.347 0.358

Change % (0–3 months) �0.86 9.02b,d 2.11 6.61 1.47 7.85 0.099 0.048

Change % (0–6 months) �1.61 12.80 1.40 8.21 0.71 8.34 0.120 0.142

Fat free mass (kg): P(T�G)¼ 0.538

Baseline 39.14 3.97 39.29 3.77 40.15 4.14 0.320

Month 3 39.00 4.01 38.85 3.74 39.66 4.15 0.494 0.469

Month 6 39.07 3.89 39.19 3.83 39.87 4.23 0.501 0.683

Change % (0–3 months) �0.32 3.23 �1.07 3.32 �1.19 3.37 0.294 0.342

Change % (0–6 months) �0.04 5.01 �0.20 3.99 �0.69 3.44 0.672 0.604

aPo0.05 by comparing with placebo group by ANOVA. bPo0.05 by comparing with placebo group by ANCOVA. cPo0.05 by comparing with Iso group by ANOVA.dPo0.05 by comparing with Iso group by ANCOVA. P(T�G): P-value for interaction between time and group by repeated-measure ANOVA. *P-value by ANOVA

analysis. **P-value by ANCOVA analysis after controlling for YSM, baseline body weight, baseline physical activity, baseline dietary energy, and soy protein intake and

their interaction terms with treatment. The P-value for a,b,c,d are from LSD, post hoc multiple comparison tests. All the mean values and percentage changes

presented in the table were unadjusted.

Soy protein and body compositionZ-m Liu et al

6

International Journal of Obesity

Despite data from animal models generally supportive for

the effects of soy protein or isoflavones on weight and fat

loss when isocaloric diets are administered,19 current

information regarding similar effects in humans are limited

with controversial results. Only three randomized placebo-

controlled trials using soy protein with or without iso-

flavones as supplements examined the effects on body

composition in peri- or postmenopausal women. One trial

reported positive findings on body fat reduction,12 whereas

the other two did not.15,16 A study by Sites et al.12 in 15

postmenopausal women reported that, compared with an

isocaloric casein placebo, a 3-month daily shake supplement

containing 20 g soy protein and 160 mg isoflavones pre-

vented an increase in subcutaneous and total abdominal fat,

whereas BW was unaffected. The study by Moeller et al.16

among perimenopausal women showed that isoflavone-rich

soy protein (40 g soy protein and 80.4 mg isoflavones) did

not affect body mass, total body fat, and lean mass compared

with whey protein or isoflavone-poor soy protein. A

12-month RCT15 conducted among older postmenopausal

women (60–75 years) reported that 25.6 g soy protein

containing 99 mg isoflavones had no effect on BMI and

WHR compared with milk protein. Differences in population

characteristics, types of soy supplements, and methods for

outcome measures may explain the inconsistent findings

among studies. Women at different menopausal stage may

respond differently to soy supplementation. Our subgroup

analysis also suggested a more pronounced soy protein effect

in early menopausal women (o4 YSM).

Our finding of a nil effect of isoflavones supplementation

without soy protein on anthropometric profile and body

composition was consistent with an earlier finding,20 but not

with two others.21,22 It has been suggested that the purified

isoflavones or isolated soy protein are not as effective as

intact soy foods on the improvement of body composition

and lipids profile.13,14,23,24 The favorable effect might be

attributed to the presence of the other components in soy

(such as soy fiber, saponin, polysaccharides, phytosterol,

unsaturated fatty acid, etc.) and their interactions.25

Compared with some notable positive findings,12,14,26 our

trial has observed a mild effect of soy on BW and BF%

reduction, whereas this magnitude was comparable to the

other two studies with the use of soy-based diet among

overweight or obese patients.13,27 However, even this mild

degree of effect may have important public health implica-

tions. Evidence has indicated that even a 1-kg increase in BW

Table 3 Years since menopause-stratified change percentage of body composition at 3- and 6-month treatment by three study arms

Change % o4 years YSM P* X4 years YSM P*

Soy group Iso group Milk group Soy group Iso group Milk group

(n¼30) (n¼33) (n¼34) (n¼30) (n¼ 27) (n¼ 26)

Mean s.d. Mean s.d. Mean s.d. Mean s.d. Mean s.d. Mean s.d.

Body weight (kg)

0–3 months �1.24 2.56a 0.25 2.17 �0.52 2.19 0.058 �0.32 2.57 �0.94 2.48 1.96 9.15 0.129

0–6 months �1.74 3.14a,b 0.40 2.30 �0.50 2.56 0.023 �0.62 2.83 �0.55 2.63 0.07 2.59 0.558

BMI (kg m�2)

0–3 months �1.24 2.56a 0.25 2.17 �0.52 2.19 0.058 �0.32 2.57 �0.94 2.48 1.96 9.15 0.100

0–6 months �1.74 3.14a,b 0.39 2.30 �0.55 2.54 0.025 �0.26 2.08 �0.55 2.63 0.07 2.59 0.592

Waist circumference (cm)

0–3 months �1.13 3.05 �0.20 2.99 �0.27 2.70 0.455 �0.23 4.29 �0.71 5.25 0.12 3.34 0.443

0–6 months �2.58 3.39a,b �0.52 3.14 �0.59 3.21 0.053 �0.30 3.81 �0.40 4.40 �0.64 4.00 0.862

Waist-to-hip ratio (WHR)

0–3 months �0.78 3.76 �0.51 2.82 0.24 3.64 0.523 �0.87 3.99 �1.00 5.71 1.03 5.91 0.319

0–6 months �1.88 3.71 �0.80 3.35 0.05 3.34 0.181 �0.92 3.74 �0.53 5.57 �0.03 5.48 0.994

Body fat %

0–3 months �1.04 7.14 2.16 5.79 0.59 6.45 0.112 0.24 7.52 1.52 5.25 3.46 6.77 0.265

0–6 months �1.46 9.86a 2.28 5.91 1.06 6.29 0.092 �2.03 9.50 0.85 10.50 2.28 7.41 0.252

Fat mass (kg)

0–3 months �1.88 9.26a 3.06 6.86 1.24 7.43 0.041 0.16 8.82 0.93 6.22 3.09 8.24 0.377

0–6 months �2.64 12.23a 3.08 7.13 1.45 7.88 0.053 �0.58 13.47 �0.65 9.07 2.23 8.83 0.579

Free fat mass (kg)

0–3 months �0.51 2.52 �0.89 3.26 �0.76 3.74 0.837 �0.12 3.85 �1.30 3.44 �1.75 2.78 0.260

0–6 months �0.81 2.98 �0.35 3.56 �0.39 3.37 0.985 0.73 6.41 �0.01 4.53 �1.09 3.55 0.235

aPo0.05 by comparing with Iso group by ANCOVA. bPo0.05 by comparing with placebo group by ANCOVA. The P-value for a,b are from LSD, post hoc multiple

comparison tests. *P-value by ANCOVA after controlling for baseline body weight, baseline physical activity, baseline dietary energy, and soy protein intake.

Soy protein and body compositionZ-m Liu et al

7

International Journal of Obesity

is associated with a 9% relative increase in diabetes

prevalence in the general population.28 For every 1% above

a desirable BMI, the risk for CHD increases by 3.3% for

women and by 3.6% for men.29 Weight loss from pharma-

cotherapy is also generally modest ranging from 2.6 to

4.5 kg, or B2–3% of initial BW after even a longer treatment

period (26–52 weeks) than ours.30 Thus, it is not expected

that soy, as a dietary supplement, will be as effective as the

weight reduction agents in improving BW. However, the

beneficial effects of soy, as part of diet, in weight control and

body fat reduction would warrant further investigation.

The mild effect of soy in our study could also possibly be

ascribed to the use of only a moderate amount of soy protein

in purified form, or the absence of restricted caloric intake

during the study. Furthermore, our study participants

consisted of women with prediabetes or early diabetes.

Earlier studies indicated that physiologic abnormalities of

diabetes may make weight loss more difficult and harder to

maintain.31 In addition, our participants had a mean BMI of

24.5 with 38% within the normal range. Thus, it is not

surprising to observe a less remarkable improvement in body

composition among the normal weight or mildly overweight

subjects. However, our subgroup analysis by baseline BMI

level did not show a more apparent effect in the higher BMI

than the lower BMI group, although the findings from

subgroup analysis should be treated with caution. Further

studies are warranted to clarify whether the soy effect can be

modified by baseline BMI status.

As in many other studies, we used milk protein to serve as

the placebo group. However, milk peptides may have active

biological properties relevant to weight loss and obesity

control.32 Thus, comparing two biologically active interven-

tions (soy and milk) could mask, an otherwise, a more

marked effect of soy. In addition, the increase in protein and

energy intake due to supplementation may have over-

shadowed a greater treatment effect.

The effect size may be related to treatment duration.

Although this study was of relatively long duration when

compared with other soy trials, a longer study may yield a

more notable treatment effect. In studies on hormone

treatment (HT) and body composition, the duration of

estrogen use was often several years.33–35 In the FELS study,36

women who had used estrogen therapy for 45 years had

significantly higher fat free mass than the non-users, whereas

little change was observed in studies of shorter duration.

The mechanisms whereby soy protein and isoflavones

exert its effects on body mass and adipose tissue are not

clearly understood. Except for estrogen-like activity,37 iso-

flavones may inhibit adipose tissue accumulation and

regulate adipogenic differentiation through numerous other

pathways such as PPARs and SREBPs,37 or acting as a tyrosine

kinase inhibitor.38 The capacity of soy protein to regulate

weight and lipid metabolism might be related to its amino

acid constitution. Soy peptide may improve the thermic

response of brown adipose tissue, promote satiation, and

alter secretion of gut hormones.39

Our study has several limitations. First, the use of

bioelectrical impedance method for body composition

measures had advantages but also limitations. Body fat

should ideally be measured by the gold standards, under-

water weighing, or dual x-ray absorptiometry, but these two

methods are generally impractical and expensive. Bio-

electrical impedance analysis (BIA) is a practical and good

alternative40,41 with established validity and reliability in the

assessment of body composition.42–44 The use of built-in

software to assess BF% with weight, height, age, and sex as

additional parameters improved the accuracy of body

composition measurement and with a precision error similar

to dual x-ray absorptiometry and underwater weighing.45

The BIA instrument used in our study has limitation in

differentiating regional fat from total FM. Further studies of

soy effect on abdominal fat change are required.

Another limitation is that equol production status has not

been assessed in our trial. Evidence has suggested that the

clinical efficacy of isoflavones in humans depends on their

ability to produce equol, a microbial metabolite of daidzein

and is more estrogenic than daidzein.46,47 As only 30–50% of

the Chinese population can produce equol after soyfood

ingestion, equol-producing status may be an important

modifier of the effects of isoflavones.

Compared with the earlier studies, this study has a

relatively large sample. Subjects’ compliance was good with

only 7.8% attrition rate. The study also had successful

blinding efficacy. Three supplements were formulated into

iso-caloric powder with similar nutrient profile allowing the

comparison of supplementation effects only between protein

variety (soy or milk) and isoflavone content (with or

without). Furthermore, dietary intake and physical activity,

the two most important modifiers of treatment effect, were

monitored and were found similar across the whole inter-

vention period.

This double-blind randomized placebo-controlled

trial revealed that 6-month daily supplementation with

15 g soy protein and 100 mg isoflavones resulted in a

mild but significant BW and fat reduction among post-

menopausal Chinese women. However, longer-term trials

are warranted to investigate the potential effects of soy, the

appropriate dosages, and the responsible components on

improvement of body composition among overweight and

obese populations.

Conflict of interest

The authors declare no conflict of interest.

Acknowledgements

We acknowledge University Grant Council of Hong Kong for

the funding support and Solbar industries Ltd (Israel) for

Soy protein and body compositionZ-m Liu et al

8

International Journal of Obesity

supplying the raw materials of isoflavones and isolated soy

protein for this study. The trial was supported by grant

from the University Grant Council with grant ID

CUHK4450/06 M and registered at clinicaltrials.gov with

reg. no. NCT00856882.

References

1 Koskova I, Petrasek R, Vondra K, Skibova J. Weight, bodycomposition and fat distribution changes of Czech women inthe different reproductive phases: a longitudinal study. PragueMed Rep 2007; 108: 226–242.

2 Misso ML, Jang C, Adams J, Tran J, Murata Y, Bell R et al.Differential expression of factors involved in fat metabolism withage and the menopause transition. Maturitas 2005; 51: 299–306.

3 Jacobsen DE, Samson MM, Kezic S, Verhaar HJ. PostmenopausalHRT and tibolone in relation to muscle strength and bodycomposition. Maturitas 2007; 58: 7–18.

4 Delibasi T, Berker D, Aydin Y, Pinar T, Ozbek M. Effects ofcombined female sex hormone replacement therapy on body fatpercentage and distribution. Adv Ther 2006; 23: 263–273.

5 Manson JE, Hsia J, Johnson KC, Rossouw JE, Assaf AR, Lasser NLet al. Estrogen plus progestin and the risk of coronary heartdisease. N Engl J Med 2003; 349: 523–534.

6 Hendrix SL, Wassertheil-Smoller S, Johnson KC, Howard BV,Kooperberg C, Rossouw JE et al. Effects of conjugated equineestrogen on stroke in the women’s health initiative. Circulation2006; 113: 2425–2434.

7 Aoyama T, Fukui K, Takamatsu K, Hashimoto Y, Yamamoto T. Soyprotein isolate and its hydrolysate reduce body fat of dietaryobese rats and genetically obese mice (yellow KK). Nutrition 2000;16: 349–354.

8 Naaz A, Yellayi S, Zakroczymski MA, Bunick D, Doerge DR,Lubahn DB et al. The soy isoflavone genistein decreases adiposedeposition in mice. Endocrinology 2003; 144: 3315–3320.

9 Nagasawa A, Fukui K, Kojima M, Kishida K, Maeda N, NagaretaniH et al. Divergent effects of soy protein diet on the expressionof adipocytokines. Biochem Biophys Res Commun 2003; 311:909–914.

10 Goodman-Gruen D, Kritz-Silverstein D. Usual dietary isoflavoneintake and body composition in postmenopausal women.Menopause 2003; 10: 427–432.

11 Yamori Y. Worldwide epidemic of obesity: hope for Japanesediets. Clin Exp Pharmacol Physiol 2004; 31 (Suppl 2): S2–S4.

12 Sites CK, Cooper BC, Toth MJ, Gastaldelli A, Arabshahi A, BarnesS. Effect of a daily supplement of soy protein on bodycomposition and insulin secretion in postmenopausal women.Fertil Steril 2007; 88: 1609–1617.

13 Liao FH, Shieh MJ, Yang SC, Lin SH, Chien YW. Effectivenessof a soy-based compared with a traditional low-calorie diet onweight loss and lipid levels in overweight adults. Nutrition 2007;23: 551–556.

14 Allison DB, Gadbury G, Schwartz LG, Murugesan R, Kraker JL,Heshka S et al. A novel soy-based meal replacement formula forweight loss among obese individuals: a randomized controlledclinical trial. Eur J Clin Nutr 2003; 57: 514–522.

15 Kok L, Kreijkamp-Kaspers S, Grobbee DE, Lampe JW, van derSchouw YT. Soy isoflavones, body composition, and physicalperformance. Maturitas 2005; 52: 102–110.

16 Moeller LE, Peterson CT, Hanson KB, Dent SB, Lewis DS, King DSet al. Isoflavone-rich soy protein prevents loss of hip lean massbut does not prevent the shift in regional fat distribution inperimenopausal women. Menopause 2003; 10: 322–331.

17 Yang Y (ed). China Food Composition 2004. Peking UniversityMedical Press: Beijing, 2004, pp 79–201.

18 Lukaski HC, Johnson PE, Bolonchuk WW, Lykken GI. Assessmentof fat-free mass using bioelectrical impedance measurements ofthe human body. Am J Clin Nutr 1985; 41: 810–817.

19 Cope MB, Erdman Jr JW, Allison DB. The potential role ofsoyfoods in weight and adiposity reduction: an evidence-basedreview. Obes Rev 2008; 9: 219–235.

20 Weickert MO, Reimann M, Otto B, Hall WL, Vafeiadou K, HallundJ et al. Soy isoflavones increase preprandial peptide YY (PYY), buthave no effect on ghrelin and body weight in healthy post-menopausal women. J Negat Results Biomed 2006; 5: 11.

21 Aubertin-Leheudre M, Lord C, Khalil A, Dionne IJ.Effect of 6 months of exercise and isoflavone supplementationon clinical cardiovascular risk factors in obese postmenopausalwomen: a randomized, double-blind study. Menopause 2007; 14:624–629.

22 Mori M, Aizawa T, Tokoro M, Miki T, Yamori Y. Soy isoflavonetablets reduce osteoporosis risk factors and obesity in middle-aged Japanese women. Clin Exp Pharmacol Physiol 2004; 31(Suppl 2): S39–S41.

23 Matthan NR, Jalbert SM, Ausman LM, Kuvin JT, Karas RH,Lichtenstein AH. Effect of soy protein from differently processedproducts on cardiovascular disease risk factors and vascularendothelial function in hypercholesterolemic subjects. Am J ClinNutr 2007; 85: 960–966.

24 Cassidy A. Factors affecting the bioavailability of soy isoflavonesin humans. J AOAC Int 2006; 89: 1182–1188.

25 Cederroth CR, Nef S. Soy, phytoestrogens and metabolism: areview. Mol Cell Endocrinol 2009; 304: 30–42.

26 Deibert P, Konig D, Schmidt-Trucksaess A, Zaenker KS, Frey I,Landmann U et al. Weight loss without losing muscle mass inpre-obese and obese subjects induced by a high-soy-protein diet.Int J Obes Relat Metab Disord 2004; 28: 1349–1352.

27 Li Z, Hong K, Saltsman P, DeShields S, Bellman M, Thames G et al.Long-term efficacy of soy-based meal replacements vs anindividualized diet plan in obese type II DM patients: relativeeffects on weight loss, metabolic parameters, and C-reactiveprotein. Eur J Clin Nutr 2005; 59: 411–418.

28 Mokdad AH, Ford ES, Bowman BA, Nelson DE, Engelgau MM,Vinicor F et al. Diabetes trends in the US: 1990–1998. DiabetesCare 2000; 23: 1278–1283.

29 Anderson JW, Konz EC. Obesity and disease management:effects of weight loss on comorbid conditions. Obes Res 2001; 9(Suppl 4): 326S–334S.

30 Norris SL, Zhang X, Avenell A, Gregg E, Schmid CH, Kim C et al.Efficacy of pharmacotherapy for weight loss in adults with type 2diabetes mellitus: a meta-analysis. Arch Intern Med 2004; 164:1395–1404.

31 Greenway F. Obesity medications and the treatment of type 2diabetes. Diabetes Technol Ther 1999; 1: 277–287.

32 Yalcin AS. Emerging therapeutic potential of whey proteins andpeptides. Curr Pharm Des 2006; 12: 1637–1643.

33 Sites CK, L’Hommedieu GD, Toth MJ, Brochu M, Cooper BC,Fairhurst PA. The effect of hormone replacement therapyon body composition, body fat distribution, and insulinsensitivity in menopausal women: a randomized, double-blind, placebo-controlled trial. J Clin Endocrinol Metab 2005; 90:2701–2707.

34 Chen Z, Bassford T, Green SB, Cauley JA, Jackson RD, LaCroix AZet al. Postmenopausal hormone therapy and body composition-Fa substudy of the estrogen plus progestin trial of the Women’sHealth Initiative. Am J Clin Nutr 2005; 82: 651–656.

35 Thorneycroft IH, Lindsay R, Pickar JH. Body composition duringtreatment with conjugated estrogens with and without medroxy-progesterone acetate: analysis of the women’s Health, Osteo-porosis, Progestin, Estrogen (HOPE) trial. Am J Obstet Gynecol2007; 197: 137 e1-7.

36 Guo SS, Zeller C, Chumlea WC, Siervogel RM. Aging, bodycomposition, and lifestyle: the Fels Longitudinal Study. Am J ClinNutr 1999; 70: 405–411.

Soy protein and body compositionZ-m Liu et al

9

International Journal of Obesity

37 Orgaard A, Jensen L. The effects of soy isoflavones on obesity. ExpBiol Med 2008; 233: 1066–1080.

38 Harmon AW, Patel YM, Harp JB. Genistein inhibits CCAAT/enhancer-binding protein beta (C/EBPbeta) activity and 3T3-L1adipogenesis by increasing C/EBP homologous protein expres-sion. Biochem J 2002; 367 (Part 1): 203–208.

39 Anderson GH, Tecimer SN, Shah D, Zafar TA. Protein source,quantity, and time of consumption determine the effect ofproteins on short-term food intake in young men. J Nutr 2004;134: 3011–3015.

40 Prentice AM, Jebb SA. Beyond body mass index. Obes Rev 2001; 2:141–147.

41 Lee SY, Gallagher D. Assessment methods in human bodycomposition. Curr Opin Clin Nutr Metab Care 2008; 11: 566–572.

42 Vasudev S, Mohan A, Mohan D, Farooq S, Raj D, Mohan V.Validation of body fat measurement by skinfolds and twobioelectric impedance methods with DEXAFthe Chennai UrbanRural Epidemiology Study [CURES-3]. J Assoc Physicians India2004; 52: 877–881.

43 Ritchie JD, Miller CK, Smiciklas-Wright H. Tanita foot-to-footbioelectrical impedance analysis system validated in older adults.J Am Diet Assoc 2005; 105: 1617–1619.

44 Boneva-Asiova Z, Boyanov MA. Body composition analysis byleg-to-leg bioelectrical impedance and dual-energy X-ray absorp-tiometry in non-obese and obese individuals. Diabetes Obes Metab2008; 10: 1012–1018.

45 Ellis KJ. Human body composition: in vivo methods. Physiol Rev2000; 80: 649–680.

46 Wu J, Oka J, Ezaki J, Ohtomo T, Ueno T, Uchiyama S et al. Possiblerole of equol status in the effects of isoflavone on bone and fatmass in postmenopausal Japanese women: a double-blind,randomized, controlled trial. Menopause 2007; 14: 866–874.

47 Muthyala RS, Ju YH, Sheng S, Williams LD, Doerge DR,Katzenellenbogen BS et al. Equol, a natural estrogenic metabolitefrom soy isoflavones: convenient preparation and resolution ofR- and S-equols and their differing binding and biological activitythrough estrogen receptors alpha and beta. Bioorg Med Chem2004; 12: 1559–1567.

Soy protein and body compositionZ-m Liu et al

10

International Journal of Obesity