Impact of soy milk consumption on cardiometabolic risk factors: A systematic review and meta-analysis of randomized controlled trials

Abstract

Background
Soy milk contains some beneficial components such as isoflavones which can exert favorable effects on the cardiovascular health. The current study aimed to comprehensively evaluate the potential effects of soy milk consumption on cardiometabolic risk factors in adults.

Methods
Relevant articles published up to June 2020 were systematically retrieved from SCOPUS, PubMed/MEDLINE, EMBASE, and Web of Science databases. In our study, we included all the randomized controlled trials (RCTs) investigating the impact of soy milk consumption on various cardiometabolic risk factors in adults (age ≥ 18 years). A meta-analysis of the eligible studies was performed using the random-effects model.

Results
The quantitative meta-analysis of 18 eligible RCTs (665 participants, age range 18–65 years) demonstrated that the consumption of soy milk significantly reduced systolic (P < 0.001) and diastolic (P = 0.002) blood pressure, total (P = 0.001) and low-density lipoprotein (P = 0.041) cholesterol, waist circumference (P = 0.005), C-reactive protein (P < 0.001), and tumor necrosis factor-alpha (P = 0.016). Significant between-study heterogeneity was found for the pooled effect sizes of blood pressure and low-density lipoprotein cholesterol. In addition, the subgroup analyses indicated that the decrease in systolic blood pressure (SBP) was more pronounced when soy milk was consumed for ≤ 4 weeks. However, there were no significant differences between soy milk and control groups for the other factors, namely body weight, body mass index (BMI), high-density lipoprotein cholesterol, triglycerides, fasting blood glucose (FBG), and fasting insulin, interleukin-6, and fibrinogen.

Conclusions
The current systematic review and meta-analysis revealed that incorporating soy milk into the diet might favorably affect several cardiometabolic risk factors in both healthy and unhealthy individuals.

Full text

Journal of Functional Foods 83 (2021) 104499
1756-4646/© 2021 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
Impact of soy milk consumption on cardiometabolic risk factors: A
systematic review and meta-analysis of randomized controlled trials
Mohammad Hassan Sohouli
a
,
b
,
1
, Abolfazl Lari
b
,
1
, Somaye Fatahi
c
,
a
, Farzad Shidfar
b
,
*
,
Mihnea-Alexandru G˘
aman
d
,
e
, Nathalia Sernizon Guimar˜
aes
f
, Ghufran Abdullatif Sindi
g
,
Rasha Abdulaziz Mandili
g
, Ghaida Rashed Alzahrani
g
, Rahaf Abdulrashid Abdulwahab
g
,
Alhanouf Mohammed Almuihi
g
, Faris Mohammed Alsobyani
h
, Amna Malik Albu Mahmud
h
,
Osama Nazzal
h
, Lama Alshaibani
h
, Shouq Elmokid
i
, Ahmed Abu-Zaid
i
a
Student Research Committee, Faculty of Public Health Branch, Iran University of Medical Sciences, Tehran, Iran
b
Department of Nutrition, School of Public Health, Iran University of Medical Sciences, Tehran, Iran
c
Pediatric Gastroenterology, Hepatology and Nutrition Research Center, Research Institute for Children’s Health, Shahid Beheshti University of Medical Sciences, Tehran,
Iran
d
Faculty of Medicine, “Carol Davila” University of Medicine and Pharmacy, Bucharest, Romania
e
Department of Hematology, Center of Hematology and Bone Marrow Transplantation, Fundeni Clinical Institute, Bucharest, Romania
f
Department of Nutrition, Federal University of Ouro Preto, Ouro Preto, Minas Gerais, Brazil
g
College of Medicine, Umm Al-Qura University, Mecca, Saudi Arabia
h
College of Medicine and Medical Sciences, Arabian Gulf University, Manama, Bahrain
i
College of Medicine, Alfaisal University, Riyadh, Saudi Arabia
ARTICLE INFO
Keywords:
Soy milk
Cardiometabolic disease
Systematic review
Meta-analysis
ABSTRACT
Background: Soy milk contains some benecial components such as isoavones which can exert favorable effects
on the cardiovascular health. The current study aimed to comprehensively evaluate the potential effects of soy
milk consumption on cardiometabolic risk factors in adults.
Methods: Relevant articles published up to June 2020 were systematically retrieved from SCOPUS, PubMed/
MEDLINE, EMBASE, and Web of Science databases. In our study, we included all the randomized controlled trials
(RCTs) investigating the impact of soy milk consumption on various cardiometabolic risk factors in adults (age ≥
18 years). A meta-analysis of the eligible studies was performed using the random-effects model.
Results: The quantitative meta-analysis of 18 eligible RCTs (665 participants, age range 18–65 years) demon-
strated that the consumption of soy milk signicantly reduced systolic (P <0.001) and diastolic (P =0.002)
blood pressure, total (P =0.001) and low-density lipoprotein (P =0.041) cholesterol, waist circumference (P =
0.005), C-reactive protein (P <0.001), and tumor necrosis factor-alpha (P =0.016). Signicant between-study
heterogeneity was found for the pooled effect sizes of blood pressure and low-density lipoprotein cholesterol. In
addition, the subgroup analyses indicated that the decrease in systolic blood pressure (SBP) was more pro-
nounced when soy milk was consumed for ≤4 weeks. However, there were no signicant differences between
soy milk and control groups for the other factors, namely body weight, body mass index (BMI), high-density
lipoprotein cholesterol, triglycerides, fasting blood glucose (FBG), and fasting insulin, interleukin-6, and
brinogen.
Conclusions: The current systematic review and meta-analysis revealed that incorporating soy milk into the diet
might favorably affect several cardiometabolic risk factors in both healthy and unhealthy individuals.
Abbreviations: DBP, diastolic blood pressure; ES, effect size; FBG, fasting blood glucose; HDL, high-density lipoprotein; LDL, low-density lipoprotein; MetS,
metabolic syndrome; SBP, systolic blood pressure; TC, total cholesterol; TG, triglyceride; T2D, type 2 diabetes; RCT, randomized controlled trials; WC, Waist
circumference.
* Corresponding author.
E-mail address: shidfar.f@iums.ac.ir (F. Shidfar).
1
Co-rst authors.
Contents lists available at ScienceDirect
Journal of Functional Foods
journal homepage: www.elsevier.com/locate/jff
https://doi.org/10.1016/j.jff.2021.104499
Received 16 November 2020; Received in revised form 8 April 2021; Accepted 15 April 2021

Journal of Functional Foods 83 (2021) 104499
2
1. Introduction
Cardiometabolic disorders affect large numbers of subjects world-
wide and are recognized as one of the leading causes of death (de
Magalh˜
aes Cunha et al., 2018; Rao, 2018). More than 60% of the deaths
from chronic kidney disease, diabetes, and cardiovascular disorders are
linked to the presence of cardiometabolic risk factors, such as dyslipi-
demia, inammation, hypertension, and obesity (Danaei et al., 2014).
Moreover, these risk factors are associated with a signicant increase in
medical expenditures as well as loss of productivity (McQueen et al.,
2016). Thus, many researchers have focused on unravelling non-
pharmacological approaches to the management of cardiometabolic
risk factors (O’Keefe, Gheewala, & O’Keefe, 2008). Indeed, lifestyle
changes involving modication of dietary habits are increasingly
emerging as key steps in the management of cardiometabolic risk factors
(Micha et al., 2017; Dariush Mozaffarian, Wilson, & Kannel, 2008).
Soy milk, which is derived from whole soybeans, is one of the most
popular functional beverages (Wang et al., 2013). Due to its high
nutritional value, soy milk is a suitable milk-substitute for vegans/veg-
etarians and those who suffer from milk allergy or lactose intolerance. It
is also regarded as a low-cost and high-quality source of protein and
energy for malnourished subjects, as well as in populations with an
insufcient supply of cow milk (Mazumder & Begum, 2016; Sethi, Tyagi,
& Anurag, 2016). Soy milk contains some benecial components such as
isoavones and polyphenols which can exert favorable effects on the
cardiovascular health (Takatsuka et al., 2000). Moreover, as a rich
source of isoavones, soy milk intake is associated with a lower inci-
dence of cancer, osteoporosis, menopausal symptoms, and cardiovas-
cular diseases (Woodside, Brennan, & Cantwell, 2016).
The potential health-promoting effects of soy milk consumption on
several cardiometabolic risk factors have been examined in several
interventional studies (Azadbakht & Nurbakhsh, 2011; Beavers, Serra,
Beavers, Cooke, & Willoughby, 2009; Eslami et al., 2019; SH Faghih,
Abadi, Hedayati, & Kimiagar, 2011). However, there are discrepancies
in the reported results. Some randomized controlled trials (RCTs) have
reported on the benecial effects of soy milk on several cardiometabolic
risk factors, that is, blood pressure, glycemic prole, and inammatory
markers in non-alcoholic fatty liver disease (NAFLD) and type 2 diabetes
mellitus patients (Maleki et al., 2019; Miraghajani, Esmaillzadeh,
Najafabadi, Mirlohi, & Azadbakht, 2012). However, several studies have
reported contradictory results (Beavers et al., 2009; Nourieh, Keshavarz,
HosseinzadehAttar, & Azadbakht, 2012).
To our knowledge, there is no systematic review and meta-analysis
regarding the impact of soy milk on cardiometabolic risk factors. Hence,
we sought to conduct a systematic review and meta-analysis of RCTs to
assess the overall impact of soy milk consumption on cardiometabolic
risk factors.
2. Methods
2.1. Search strategy
The current systematic review was executed in accordance with the
Preferred Reporting Items for Systematic Reviews and Meta-Analyses
(PRISMA) statement criteria (Moher et al., 2015). The study protocol
has been previously registered with the PROSPERO database (ID num-
ber: 219486). The relevant articles were retrieved from four online
databases: PubMed/MEDLINE, SCOPUS, Web of Science, and EMBASE.
The literature was systematically searched from February 1994 until
July 2020 using the following Medical Subject Headings (MeSH) and
text keywords: (“Soya Milk” OR “Soy Milk” OR “Soy drink” OR “Soy
beverage”) AND (“RCT” OR “Intervention” OR “Trial” OR “Control*” OR
“Clinical” OR “Random*” OR “Placebo” OR “Assignment” OR “Alloca-
tion”). No language limitation was imposed to the literature search.
Moreover, we hand-searched the references of any identied review
papers to detect other potentially relevant articles.
2.2. Study selection
The studies were selected for full-text review using the EndNote
software if they met the following PICOS evidence-based criteria: 1)
Patients: adult male or female participants (aged ≥18 years), 2) Inter-
vention: soy milk consumption, 3) Comparator: placebo or cow milk, 4)
Outcomes: reported sufcient data on at least one of the outcomes of
interest, namely body weight, body mass index (BMI), waist circum-
ferences (WC), systolic blood pressure (SBP), diastolic blood pressure
(DBP), total cholesterol (TC), low-density lipoprotein cholesterol (LDL-
C), high-density lipoprotein cholesterol (HDL-C), triglycerides (TG),
fasting blood glucose (FBG), insulin, C-reactive protein (CRP), tumor
necrosis factor-
α
(TNF-
α
), interleukin 6 (IL-6), and brinogen, and 5)
Study design: RCTs. We excluded non-randomized trials, studies without
appropriate control groups, animal studies, review articles, unpublished
studies (inclusive of conference abstracts and high quality RCTs), and
studies that administered soy milk in combination with other com-
pounds such as phytosterols. The primary outcomes of this study
included TC, LDL-C, HDL-C, TG, SBP, DBP, WC, FBG, and CRP concen-
trations. Body weight, BMI, insulin, TNF-
α
, IL-6, and brinogen were
considered as secondary outcomes.
2.3. Data extraction
Eligible studies were independently reviewed by two authors (Mh.S.
and A.L.) and the following data were extracted: rst author’s name,
study location, publication year, RCT design (cross-over or parallel),
sample size (intervention and control groups) participant characteristics
(gender, age, and health status), duration of intervention, the amount of
soy milk consumed, and the means and standard deviations (SDs) of the
intended outcomes at baseline, post-intervention and/or changes be-
tween baseline and post-intervention.
2.4. Quality assessment
The details of the quality assessment of the included studies are
presented in Table 2. The quality of the RCTs was methodologically
evaluated based on the Cochrane risk of bias criteria (JPT Higgins &
Wells, 2011). Two authors independently (S.F. and Mh.S.) rated each
study as having a low, high, or unclear risk of bias based on the following
potential sources of bias: blinding of outcome assessment; allocation
concealment; blinding of participants and personnel; random sequence
generation; incomplete outcome data; selective reporting; and other
bias. Any discrepancies were resolved via discussion with a third author
(F.S.).
2.5. Statistical analysis
Data were analyzed using STATA version 12.0 software. Standard
formulas were applied to convert different formats of data to the mean
and standard deviations (SDs) (JP Higgins, 2011; Hozo, Djulbegovic, &
Hozo, 2005). For instance, in the absence of SDs of the change, we
calculated it through the following formula: SD changes =square root
[(SD baseline
2
+SD nal
2
) - (2 ×R ×SD baseline ×SD nal)]. Also, we
converted the standard error of the mean (SEM) to the SD using the
following formula: SD =SEM ×√n, where “n” is the number of subjects
in each group. Heterogeneity among studies was appraised using the I-
squared (I
2
) statistic (J. P. Higgins, Thompson, Deeks, & Altman, 2003).
The random-effects model was used for meta-analysis of study outcomes.
The weighting of studies was done using the generic inverse variance
method. In case of multiple evaluations in a single study group, the
values belonging to the longest time point were used for the analyses.
The effect size was expressed as weighed mean difference (WMD) and
95% condence interval (CI). Moreover, in order to determine the po-
tential sources of heterogeneity, subgroup analyses based on dose,
country, and duration of intervention was performed. The sensitivity
M. Hassan Sohouli et al.

Journal of Functional Foods 83 (2021) 104499
3
analysis was also done to assess the impact of every individual study on
the pooled effect size. The publication bias was examined using the
formal Egger’s test (Egger, Smith, Schneider, & Minder, 1997)
3. Results
3.1. Study selection
Fig. 1 depicts the literature search and selection process. In the pri-
mary systematic search, 865 studies were detected. After removing the
duplicate records, a total of 475 publications remained. Thereafter, two
investigators (Mh.S. and S.F.) independently and blindly screened the
titles/abstracts of the retrieved studies and excluded 446 articles which
did not meet the inclusion criteria. After secondary review of full-texts
(Mh.S. and S.F.), 29 studies remained of which 11 were excluded due
to different reasons. Finally, 18 studies (Azadbakht & Nurbakhsh, 2011;
Beavers et al., 2009; Eslami et al., 2019; SH Faghih et al., 2011; S Faghih,
Hedayati, Abadi, & Kimiagar, 2009; Gardner, Messina, Kiazand, Morris,
& Franke, 2007; Keshavarz, Nourieh, Attar, & Azadbakht, 2012;
Lukaszuk, Luebbers, & Gordon, 2007; Maleki et al., 2019; Miraghajani
et al., 2012; Miraghajani et al., 2013; Mitchell & Collins, 1999;
Mohammad-Shahi, Mowla, Haidari, Zarei, & Choghakhori, 2016;
Nourieh et al., 2012; ¨
Onning, Åkesson, ¨
Oste, & Lundquist, 1998;
Onuegbu, Olisekodiaka, Onibon, Adesiyan, & Igbeneghu, 2011; Rivas,
Garay, Escanero, Cia Jr, et al., 2002; Takatsuka et al., 2000) with 19
treatment arms were eligible for inclusion in the present meta-analysis.
Five studies provided data for FBG (Gardner et al., 2007; Keshavarz
et al., 2012; Maleki et al., 2019; Miraghajani et al., 2013; ¨
Onning et al.,
1998), four articles for insulin (29,30,32, 35), seven trials for body
weight (Azadbakht & Nurbakhsh, 2011; Eslami et al., 2019; SH Faghih
et al., 2011; Keshavarz et al., 2012; Lukaszuk et al., 2007; Miraghajani
et al., 2013; Mohammad-Shahi et al., 2016), six studies for BMI (Azad-
bakht & Nurbakhsh, 2011; Eslami et al., 2019; S Faghih et al., 2009;
Keshavarz et al., 2012; Lukaszuk et al., 2007; Mohammad-Shahi et al.,
2016), ve trials for WC (Azadbakht & Nurbakhsh, 2011; Eslami et al.,
2019; S Faghih et al., 2009; Keshavarz et al., 2012; Lukaszuk et al.,
2007), ve articles for blood pressure (Azadbakht & Nurbakhsh, 2011;
Keshavarz et al., 2012; Maleki et al., 2019; Miraghajani et al., 2013;
Rivas, Garay, Escanero, Cia Jr, et al., 2002), four studies for CRP (Eslami
et al., 2019; Miraghajani et al., 2012; Mohammad-Shahi et al., 2016;
Nourieh et al., 2012) and IL-6 (Beavers et al., 2009; Miraghajani et al.,
2012; Mohammad-Shahi et al., 2016; Nourieh et al., 2012), and three
trials for TNF-
α
(Beavers et al., 2009; Miraghajani et al., 2012;
Mohammad-Shahi et al., 2016) and brinogen (Keshavarz et al., 2012;
Maleki et al., 2019; Miraghajani et al., 2012). Eight trials reported the
effect of soy milk on TG (Eslami et al., 2019; Gardner et al., 2007;
Miraghajani et al., 2013; Mitchell & Collins, 1999; Nourieh et al., 2012;
¨
Onning et al., 1998; Onuegbu et al., 2011; Takatsuka et al., 2000), and
seven trials reported the data on LDL-C, HDL-C, and TC (Eslami et al.,
2019; Gardner et al., 2007; Miraghajani et al., 2013; Mitchell & Collins,
1999; Nourieh et al., 2012; ¨
Onning et al., 1998; Onuegbu et al., 2011;
Takatsuka et al., 2000).
3.2. Study characteristics
The characteristics of the eligible RCTs are displayed in Table 1. Ten
studies were conducted in Iran (Azadbakht & Nurbakhsh, 2011; Eslami
et al., 2019; SH Faghih et al., 2011; S Faghih et al., 2009; Keshavarz
et al., 2012; Maleki et al., 2019; Miraghajani et al., 2012; Miraghajani
et al., 2013; Mohammad-Shahi et al., 2016; Nourieh et al., 2012) and
Records identified through database
searching
(n =865)
Screening
Included
Eligibility
Identification
Additional records identified
through other sources
(n = 3)
Records remaining after the removal of
duplicates
(n =475)
Records screened
(n = 475)
Records excluded
(n = 446)
Full-text articles assessed
for eligibility
(n = 29)
11 articles excluded
Non-RCT, n=4
Lack of control groups, n=2
Evaluating the effect of a soy milk
along with other components, N=1
Full-text not available, n=2
No outcomes of interest=2
18 RCTs included in the
systematic review and
meta-analysis
(n =18)
Fig. 1. Flow diagram demonstrating the study selection process for the meta-analysis (PRISMA [Preferred Reporting Items for Systematic Reviews and Meta-
Analyses] diagram). RCTs, randomized controlled trials.
M. Hassan Sohouli et al.

Journal of Functional Foods 83 (2021) 104499
4
Table 1
Characteristics of the randomized controlled trials included in this meta-analysis.
Study ID Country Study
design
Participants/ Sex Sample size
Intervention/
control
Mean
age or
range
age
Intervention
group
(components
milk or diet)
Control group
(components
milk or diet)
Measure or
range of
Intervention
group
Duration
(weeks)
Outcomes
Keshavarz et al
2012
Iran Cross-
over
Overweight and obese
Subjects/Female
24/24 37.7 soy milk +
reduced 200
to 500 kcal/
day (CHO:
3.5, Fat: 1 ,
Pro: 2.5 g/
100 ml)
CHO:
50–60%
Fat: <30%
Pro: 15–20%
cow’s milk +
reduced 200
to 500 kcal/
day (CHO:
4.9, Fat: 1.5 ,
Pro: 3.3 g/
100 ml)
CHO:
50–60%
Fat: <30%
Pro: 15–20%
240 ml 4 Fasting
blood sugar
Fasting
insulin
Systolic
blood
pressure
Diastolic
blood
pressure
Weight
BMI
Fibrinogen
WC
Malekiet al. 2019 Iran Parallel NAFLD patients/Male
and Female
31/31 45.89 soy milk +
500-kcal
decit diet
(CHO: 9.15,
Fat: 4 , Pro:
6.75 g /per
serving)
CHO: 55%
Fat: 30%
Pro: 15%
500-kcal
decit diet:
CHO: 55%
Fat: 30%
Pro: 15%
240 ml 8 Fasting
blood sugar
Fasting
insulin
Triglycerides
HDL
cholesterol
LDL
cholesterol
Total
cholesterol
Systolic
blood
pressure
Diastolic
blood
pressure
Fibrinogen
Miraghajaniet al.
2012 and 2013
Iran Cross-
over
Diabetic nephropathy
patients/ Male and
Female
25/25 51 soy milk
(CHO: 3.5,
Fat: 1, Pro:
2.5 per 100 g)
cow’s milk
(CHO: 4.9,
Fat: 1.5, Pro:
3.3 per 100 g)
240 ml 4 Fasting
blood sugar
Fasting
insulin
Triglycerides
HDL
cholesterol
LDL
cholesterol
Total
cholesterol
Systolic
blood
pressure
Diastolic
blood
pressure
Weight
CRP
TNF-
α
IL-6
Fibrinogen
Gardner et al.,
2007
USA Parallel Hypercholesterolemic
adults/ Male and
Female
12/12 52 1)soy milk
(CHO: 39,
Fat: 9.5, Pro:
25 Amount
Consumed/
Day)
2) whole soy
milk (CHO:
40, Fat: 13.2,
Pro: 25
Amount
Consumed/
Day)
dairy milk
(CHO: 35,
Fat: 5.8,
Pro:25
Amount
Consumed/
Day)
828 ml 4 Fasting
blood sugar
Fasting
insulin
HDL
cholesterol
LDL
cholesterol
Total
cholesterol
Systolic
blood
pressure
Diastolic
blood
pressure
¨
Onning et al 1998 Sweden Parallel 12/12 31.7 875 ml 4
(continued on next page)
M. Hassan Sohouli et al.

Journal of Functional Foods 83 (2021) 104499
5
Table 1 (continued )
Study ID Country Study
design
Participants/ Sex Sample size
Intervention/
control
Mean
age or
range
age
Intervention
group
(components
milk or diet)
Control group
(components
milk or diet)
Measure or
range of
Intervention
group
Duration
(weeks)
Outcomes
Healthy subjects/
Male and Female
soy milk
(CHO: 40,
Fat: 20, Pro:
30 per liter)
cow’s milk
(CHO: 49,
Fat: 17,
Pro:34 per
liter)
Fasting
blood sugar
Fasting
insulin
Triglycerides
HDL
cholesterol
LDL
cholesterol
Total
cholesterol
Systolic
blood
pressure
Diastolic
blood
pressure
Weight
Nourieh et al
2012
Iran Cross-
over
overweight and obese
subjects/ Female
24/24 37.7 soy milk +
reduced 200
to 500 kcal/
day (Fat: 1 g/
100 ml)
CHO:
50–60%
Fat: <30%
Pro: 15–20%
cow’s milk +
reduced 200
to 500 kcal/
day (Fat: 1.5
g/100 ml)
CHO:
50–60%
Fat: <30%
Pro: 15–20%
240 ml 4 Triglycerides
HDL
cholesterol
LDL
cholesterol
Total
cholesterol
Systolic
blood
pressure
Diastolic
blood
pressure
CRP
IL-6
Eslami et al.,
2019
Iran Parallel NAFLD patients/
Female
24/24 45.7 soy milk +
500-kcal
decit diet
(CHO: 9.15,
Fat: 4 , Pro:
6.75 g /per
serving)
CHO: 55%
Fat: 30%
Pro: 15%
500-kcal
decit diet:
CHO: 55%
Fat: 30%
Pro: 15%
240 ml 8 Triglycerides
HDL
cholesterol
LDL
cholesterol
Total
cholesterol
Systolic
blood
pressure
Diastolic
blood
pressure
CRP
Weight
BMI
WC
Azadbakht et al
2011
Iran Cross
over
overweight and obese
subjects/ Female
23/23 22.2 soy milk +
reduced 200
to 500 kcal/
day (CHO:
3.5, Fat: 1 ,
Pro: 2 g/100
ml)
CHO:
50–60%
Fat: <30%
Pro: 15–20%
cow’s milk +
reduced 200
to 500 kcal/
day
CHO:
50–60%
Fat: <30%
Pro: 15–20%
240 ml 6 Systolic
blood
pressure
Diastolic
blood
pressure
Weight
BMI
WC
Mohammad-
Shahi et al.,
2016
Iran Cross-
over
rheumatoid arthritis
patients/ Female
24/24 45.72 soy milk
(CHO: 7.5,
Fat: 1, Pro:
2.4 per 100 g)
cow’s milk
(CHO: 4.9,
Fat: 1.5, Pro:
3.3 per 100 g)
240 ml 4 Fasting
blood sugar
Fasting
insulin
HOMA-IR
Triglycerides
HDL
cholesterol
LDL
cholesterol
Total
cholesterol
(continued on next page)
M. Hassan Sohouli et al.

Journal of Functional Foods 83 (2021) 104499
6
Table 1 (continued )
Study ID Country Study
design
Participants/ Sex Sample size
Intervention/
control
Mean
age or
range
age
Intervention
group
(components
milk or diet)
Control group
(components
milk or diet)
Measure or
range of
Intervention
group
Duration
(weeks)
Outcomes
Systolic
blood
pressure
Diastolic
blood
pressure
CRP
TNF-
α
IL-6
Weight
BMI
Faghih et al 2009
and 2011
Iran Parallel overweight and obese
subjects/ Female
21/20 35 soy milk(low
fat milk
1.5%) +500-
kcal decit
diet:
CHO: 55%
Fat: 27%
Pro: 18%
500-kcal
decit diet:
CHO: 55%
Fat: 27%
Pro: 18%
720 ml 8 Fasting
blood sugar
Fasting
insulin
HOMA-IR
Triglycerides
HDL
cholesterol
LDL
cholesterol
Total
cholesterol
Systolic
blood
pressure
Diastolic
blood
pressure
Weight
BMI
WC
Beavers et al 2009 USA Parallel Healthy
postmenopausal
women/ Female
16/15 50 soy milk
(CHO: 19,
Fat: 4 , Pro: 6
g /per
serving)
dairy milk
(CHO: 12,
Fat: 4.5 , Pro:
8 g /per
serving)
740 ml 4 TNF-
α
IL-6
Takatsuka et al
2000
Japan Parallel Healthy
premenopausal
women/ Female
27/25 26.5 soy milk
(CHO: N/A,
Fat: 2.3 , Pro:
4.3 g /per
serving)
Usual diet 400 ml 8 Triglycerides
HDL
cholesterol
LDL
cholesterol
Total
cholesterol
Onuegbu et al
2011
Nigeria Parallel Healthy subjects/
Male and Female
42/40 20–35 soy milk
(CHO: N/A,
Fat: 2.3 , Pro:
4.3 g /per
serving)
Did not
consume
soymilk
500 ml 3 Triglycerides
HDL
cholesterol
LDL
cholesterol
Total
cholesterol
Mitchell et al
1999
UK Parallel Healthy subjects/
Male
4/3 20–50 soy milk dairy milk 1000 ml 4 Triglycerides
Total
cholesterol
Rivas et al 2002 Spain Parallel Hypertensive
patients/ Male and
Female
20/20 N/A soy milk
(CHO: 25,
Fat: 10.5, Pro:
18 gr/liter)
cow’s milk
(CHO: 13.5,
Fat: 1.5, Pro:
15.5 gr/liter)
1000 ml 12 Systolic
blood
pressure
Diastolic
blood
pressure
Lukaszuk et al
2007
USA Parallel Overweight/ Female 7/7 31.5 soy milk +
500-kcal
decit diet
skimmed
milk +500-
kcal decit
diet
720 ml 8 Weight
BMI
WC
NAFLD, non-alcoholic fatty liver disease. Pro, protein. CHO, carbohydrates. LDL, low-density lipoprotein. HDL, high-density lipoprotein. CRP, C-reactive protein. TNF-
α
, tumour necrosis factor
α
. IL-6, interleukin 6. BMI, body mass index. WC, waist circumference. HOMA-IR, Homeostatic Model Assessment of Insulin Resistance.
M. Hassan Sohouli et al.

Journal of Functional Foods 83 (2021) 104499
7
three studies in the United States of America (USA) (Beavers et al., 2009;
Gardner et al., 2007; Lukaszuk et al., 2007). The rest of the studies were
carried out in Sweden ( ¨
Onning et al., 1998), Japan (Takatsuka et al.,
2000), Nigeria (Onuegbu et al., 2011), United Kingdom (Mitchell &
Collins, 1999), and Spain (Rivas, Garay, Escanero, Cia Jr, et al., 2002).
These studies were published between the years 1998–2019. The age of
the participants ranged from 18 to 65 years. Respectively, ten and one
study were performed on women and men, and the rest on both sexes.
Twelve RCTs had a parallel design (Beavers et al., 2009; Eslami et al.,
2019; SH Faghih et al., 2011; S Faghih et al., 2009; Gardner et al., 2007;
Lukaszuk et al., 2007; Maleki et al., 2019; Mitchell & Collins, 1999;
¨
Onning et al., 1998; Onuegbu et al., 2011; Rivas, Garay, Escanero, Cia
Jr, et al., 2002; Takatsuka et al., 2000), whereas the remaining had a
cross-over design (Azadbakht & Nurbakhsh, 2011; Keshavarz et al.,
2012; Miraghajani et al., 2012; Miraghajani et al., 2013; Mohammad-
Shahi et al., 2016; Nourieh et al., 2012). The duration of the intervention
varied from 4 to 8 weeks and the amount of soy milk administered
ranged between 240 and 1000 ml/day among the analyzed studies.
Eight trials administered soy milk with a reduced-calorie (200–500 kcal)
diet (Azadbakht & Nurbakhsh, 2011; Eslami et al., 2019; SH Faghih
et al., 2011; S Faghih et al., 2009; Keshavarz et al., 2012; Lukaszuk et al.,
2007; Maleki et al., 2019; Nourieh et al., 2012). Six studies were con-
ducted on healthy overweight/obese individuals (Azadbakht & Nur-
bakhsh, 2011; SH Faghih et al., 2011; S Faghih et al., 2009; Keshavarz
et al., 2012; Lukaszuk et al., 2007; Nourieh et al., 2012), and the other
studies enrolled patients with non-alcoholic fatty liver disease (NAFLD)
(Eslami et al., 2019; Maleki et al., 2019), diabetic nephropathy (Mir-
aghajani et al., 2012; Miraghajani et al., 2013), hypercholesterolemia
(Gardner et al., 2007), rheumatoid arthritis (Mohammad-Shahi et al.,
2016), or hypertension (Rivas, Garay, Escanero, Cia Jr, et al., 2002). In
addition, ve RCTs enrolled healthy individuals (Beavers et al., 2009;
Mitchell & Collins, 1999; ¨
Onning et al., 1998; Onuegbu et al., 2011;
Takatsuka et al., 2000).
3.3. Meta-analysis
3.3.1. The effect of soy milk on fasting blood glucose and fasting insulin
Based on a random-effects statistical model, the meta-analysis of the
RCTs detected no signicant impact of soy milk on FBG (weighted mean
difference, WMD: −0.15 mg/dL, 95% CI: −2.19, 1.89, P =0.88) and
fasting insulin (WMD: 0.61
μ
U/mL, 95% CI: −1.35, 2.56, P =0.54). No
evidence of heterogeneity was observed among the analyzed studies for
FBG (Cochran Q test, P =0.97, I
2
=0.0%). However, the between-study
heterogeneity for fasting insulin was signicantly high (Cochran Q test,
P =0.014, I
2
=71.8%) (Fig. 2). In the subgroup analysis, the dose and
the duration of the intervention were considered as heterogeneity fac-
tors on the overall effect size for fasting insulin (Supplementary
Figure 1). As the number of articles in each subgroup was small, the
results were not reportable.
3.3.2. The effect of soy milk on the lipid prole
The pooled results from the random-effects model revealed that soy
milk consumption resulted in a signicant reduction of TC (WMD:
−8.97 mg/dL, 95% CI: −14.29, −3.65, P =0.001) and LDL-C (WMD:
−9.30 mg/dL, 95% CI: −18.20, −0.40, P =0.041) concentration.
However, soy milk intake did not reduce serum TG (WMD: −1.60 mg/
dL, 95% CI: −14.15, 10.94, P =0.80) and also did not affect serum HDL-
C (WMD: 1.43 mg/dL, 95% CI: −2.07, 4.94, P =0.42) concentration. No
signicant heterogeneity was observed between these trials for TC
(Cochran Q test, P =0.73, I
2
=0.0%) and HDL-C (Cochran Q test, P =
0.07, I
2
=50.9%). However, a signicant heterogeneity was seen for
LDL-C (Cochran Q test, P =0.008, I
2
=65.3%) and TG (Cochran Q test,
P =0.008, I
2
=65.2%) (Fig. 3). In the subgroup analysis for LDL-C and
TG, we found that the duration and the dose of the intervention could
explain this heterogeneity. However, the subgroup analysis based on
these factors did not show a signicant effect of soy milk consumption
on TG, HDL-C or LDL-C (Supplementary Figure 2, 3, 4).
3.3.3. The effect of soy milk on body composition
The meta-analysis of seven eligible RCTs for body weight and of six
RCTs for BMI showed no signicant difference in body weight (WMD:
−0.74 kg, 95% CI: −1.65, 0.17, P =0.11) and BMI (WMD: −0.21 kg/m
2
,
95% CI: −0.51, 0.09, P =0.176) following the consumption of soy milk.
However, the quantitative meta-analysis displayed that soy milk lowered
WC in a signicant manner compared with the control group (WMD:
−1.61 cm, 95% CI: −2.74, −0.48, P =0.005). There was no evidence of
signicant between-study heterogeneity (Cochran Q test, P =1.00, I
2
=
0.0% for weight; Cochran Q test, P =1.00, I
2
=0.0% for BMI; Cochran Q
test, P =0.84, I
2
=0.0% for WC) in the current meta-analysis (Fig. 4).
3.3.4. The effect of soy milk on SBP and DBP
The pooled effect sizes from ve RCTs indicated that soy milk
signicantly decreased SBP (WMD: −7.38 mmHg; 95% CI: −10.87,
−3.88, P <0.001) and DBP (WMD: −4.36 mmHg; 95% CI: −7.06,
−1.66, P =0.002). However, for SBP and DBP, the detected heteroge-
neity was signicantly high (Cochran Q test, P <0.001, I
2
=89.9% for
Table 2
Risk of bias assessment according to the Cochrane collaboration’s risk of bias assessment tool.
Study, Year (reference) Random
sequence
generation
Allocation
concealment
Blinding of
participantsand
personnel
Blinding of
outcome
assessment
Incomplete
outcome data
Selective
reporting
Overallassessment of
risk of bias
Keshavarz et al 2012 Unclear Unclear Unclear Unclear Low Low Unclear
Malekiet al. 2019 Low Unclear Unclear Low Low Low Unclear
Miraghajaniet al. 2012
and 2013
Unclear Unclear Unclear Unclear Unclear Unclear Unclear
Gardner et al., 2007 Low Unclear Unclear Unclear Low Low Unclear
¨
Onning et al 1998 Low Unclear Low Low Low Unclear Unclear
Nourieh et al 2012 Low Low Unclear Low Low Low Unclear
Eslami et al., 2019 Unclear Unclear Unclear Unclear Low Low Unclear
Azadbakht et al 2011 Unclear Unclear Unclear Unclear Unclear Unclear Unclear
Mohammad-Shahi
et al., 2016
Low Low Unclear Low Unclear Low Unclear
Faghih et al 2009 and
2011
Low Unclear Unclear Unclear Unclear Unclear Unclear
Beavers et al 2009 Low Low Low Low Low Low Low
Takatsuka et al 2000 Low Unclear Unclear Unclear Low Low Unclear
Onuegbu et al 2011 Low Unclear Low Low Low Unclear Unclear
Mitchell et al 1999 Low Low Unclear Low Low Low Unclear
Rivas et al 2002 Low Low Unclear Low Low Low Unclear
Lukaszuk et al 2007 Unclear Unclear Unclear Unclear Unclear Unclear Unclear
M. Hassan Sohouli et al.

Journal of Functional Foods 83 (2021) 104499
8
SBP and Cochran Q test, P <0.001, I
2
=89.0% for DBP), respectively
(Fig. 5). The health status of the participants (healthy versus unhealthy)
was considered as the heterogeneity factor on the overall effect size for
blood pressure. When the studies were categorized based on their du-
rations, soy milk decreased SBP in a more notable manner when it was
administered for ≤4 weeks (WMD: −8.79 mmHg, 95% CI: −16.80,
−0.78). However, no signicant effect of soy milk on DBP was observed
in the subgroup analysis based on the duration of the intervention.
(Supplementary Figure 5, 6).
3.3.5. Effect of soy milk on CRP, IL-6, TNF-
α
and brinogen
Four and three studies reported data for serum CRP and TNF-
α
as
outcome measures, respectively. The results from our meta-analysis
indicate a signicant reduction of CRP (WMD: −1.07, mg/L, 95% CI:
a) b)
NOTE: Weights are from random effects analysis
Overall (I-squared = 0.0%, p = 0.907)
ID
Maleki (2019)
Onning (1998)
Keshavarz (2012)
Gardner (a) (2007)
Miraghajani (2013)
Gardner (b) (2007)
Study
-0.15 (-2.19, 1.89)
WMD (95% CI)
-1.81 (-7.21, 3.59)
0.00 (-4.88, 4.88)
-1.33 (-5.58, 2.92)
0.00 (-4.86, 4.86)
0.00 (-7.99, 7.99)
2.00 (-2.48, 6.48)
100.00
Weight
14.32
17.53
23.13
17.67
6.53
20.82
%
-0.15 (-2.19, 1.89)
WMD (95% CI)
-1.81 (-7.21, 3.59)
0.00 (-4.88, 4.88)
-1.33 (-5.58, 2.92)
0.00 (-4.86, 4.86)
0.00 (-7.99, 7.99)
2.00 (-2.48, 6.48)
100.00
Weight
14.32
17.53
23.13
17.67
6.53
20.82
%
0-7.99 0 7.99
NOTE: Weights are from random effects analysis
Overall (I-squared = 71.8%, p = 0.014)
Study
Onning (1998)
Miraghajani (2013)
Keshavarz (2012)
ID
Maleki (2019)
0.61 (-1.35, 2.56)
2.43 (1.23, 3.63)
0.36 (-1.74, 2.46)
0.67 (-2.77, 4.11)
WMD (95% CI)
-1.54 (-3.75, 0.67)
100.00
%
32.16
25.71
17.19
Weight
24.94
0.61 (-1.35, 2.56)
2.43 (1.23, 3.63)
0.36 (-1.74, 2.46)
0.67 (-2.77, 4.11)
WMD (95% CI)
-1.54 (-3.75, 0.67)
100.00
%
32.16
25.71
17.19
Weight
24.94
0-4.11 0 4.11
Fig. 2. Forest plots from the meta-analysis of clinical trials investigating the effects of soy milk supplementation on (a) fasting blood glucose and (b) fasting insulin.
WMD: weighted mean difference. CI: condence intervals.
a) b)
c) d)
NOTE: Weights are from random effects analysis
Overall (I-squared = 0.0%, p = 0.733)
Takatsuka (2000)
ID
Nourieh (2012)
Mitchell (1999)
Miraghajani (2013)
Onning (1998)
Onuegbu (2011)
Study
Beavers (2010)
Eslami (2018)
-8.97 (-14.29, -3.65)
-11.90 (-25.00, 1.20)
WMD (95% CI)
-11.04 (-32.71, 10.63)
-20.46 (-47.20, 6.28)
-9.72 (-38.44, 19.00)
-3.86 (-14.10, 6.38)
-15.44 (-26.32, -4.56)
-0.99 (-24.48, 22.50)
-2.86 (-17.14, 11.42)
100.00
16.52
Weight
6.03
3.96
3.44
27.03
23.97
%
5.14
13.90
-8.97 (-14.29, -3.65)
-11.90 (-25.00, 1.20)
WMD (95% CI)
-11.04 (-32.71, 10.63)
-20.46 (-47.20, 6.28)
-9.72 (-38.44, 19.00)
-3.86 (-14.10, 6.38)
-15.44 (-26.32, -4.56)
-0.99 (-24.48, 22.50)
-2.86 (-17.14, 11.42)
100.00
16.52
Weight
6.03
3.96
3.44
27.03
23.97
%
5.14
13.90
0-47.2 0 47. 2
NOTE: Weights are from random effects analysis
Overall (I-squared = 65.3%, p = 0.008)
Eslami (2018)
Beavers (2010)
Onuegbu (2011)
Miraghajani (2013)
ID
Study
Onning (1998)
Nourieh (2012)
Takatsuka (2000)
-9.30 (-18.20, -0.40)
-4.49 (-18.05, 9.07)
-2.06 (-21.82, 17.70)
-30.89 (-42.22, -19.56)
1.48 (-17.35, 20.31)
WMD (95% CI)
-3.86 (-14.10, 6.38)
-11.75 (-28.25, 4.75)
-7.80 (-19.50, 3.90)
100.00
14.87
10.72
16.61
11.26
Weight
%
17.47
12.76
16.31
-9.30 (-18.20, -0.40)
-4.49 (-18.05, 9.07)
-2.06 (-21.82, 17.70)
-30.89 (-42.22, -19.56)
1.48 (-17.35, 20.31)
WMD (95% CI)
-3.86 (-14.10, 6.38)
-11.75 (-28.25, 4.75)
-7.80 (-19.50, 3.90)
100.00
14.87
10.72
16.61
11.26
Weight
%
17.47
12.76
16.31
0-42.2 0 42.2
NOTE: Weights are from random effects analysis
Overall (I-squared = 50.9%, p = 0.070)
Takatsuka (2000)
Onuegbu (2011)
Miraghajani (2013)
Nourieh (2012)
Eslami (2018)
Beavers (2010)
ID
Study
1.43 (-2.07, 4.94)
-4.00 (-9.69, 1.69)
11.58 (3.46, 19.70)
0.96 (-4.86, 6.78)
0.75 (-5.87, 7.37)
2.25 (-0.87, 5.37)
-1.69 (-11.83, 8.45)
WMD (95% CI)
100.00
18.19
12.15
17.79
15.55
27.42
8.90
Weight
%
1.43 (-2.07, 4.94)
-4.00 (-9.69, 1.69)
11.58 (3.46, 19.70)
0.96 (-4.86, 6.78)
0.75 (-5.87, 7.37)
2.25 (-0.87, 5.37)
-1.69 (-11.83, 8.45)
WMD (95% CI)
100.00
18.19
12.15
17.79
15.55
27.42
8.90
Weight
%
0-19.7 0 19.7
NOTE: Weights are from random effects analysis
Overall (I-squared = 65.2%, p = 0.008)
Miraghajani (2013)
Onuegbu (2011)
Study
Beavers (2010)
Eslami (2018)
ID
Mitchell (1999)
Nourieh (2012)
Takatsuka (2000)
-1.60 (-14.15, 10.94)
-38.60 (-73.67, -3.53)
-8.85 (-18.34, 0.64)
3.31 (-38.41, 45.03)
-3.06 (-19.49, 13.37)
WMD (95% CI)
30.98 (11.46, 50.50)
-5.26 (-24.96, 14.44)
-0.70 (-19.89, 18.49)
100.00
8.41
21.49
%
6.61
17.26
Weight
15.38
15.28
15.57
-1.60 (-14.15, 10.94)
-38.60 (-73.67, -3.53)
-8.85 (-18.34, 0.64)
3.31 (-38.41, 45.03)
-3.06 (-19.49, 13.37)
WMD (95% CI)
30.98 (11.46, 50.50)
-5.26 (-24.96, 14.44)
-0.70 (-19.89, 18.49)
100.00
8.41
21.49
%
6.61
17.26
Weight
15.38
15.28
15.57
0
-73.7 0 73. 7
Fig. 3. Forest plots from the meta-analysis of clinical trials investigating the effects of soy milk supplementation on (a) total cholesterol, (b) LDL-C, c) HDL-C and d)
TG. WMD: weighted mean difference. CI: condence intervals. LDL-C: low-density lipoprotein cholesterol. HDL-C: high-density lipoprotein cholesterol. TG:
triglycerides.
M. Hassan Sohouli et al.

Journal of Functional Foods 83 (2021) 104499
9
−1.13, −1.01, P <0.001) and TNF-
α
(WMD: −0.30, pg/mL, 95% CI:
−0.55, −0.06, P =0.016) levels following soy milk consumption. In
addition, a signicant heterogeneity was not noted among the analyzed
studies for CRP (Cochran Q test, P =0.93, I
2
=0.0%) and TNF-
α
(Cochran Q test, P =0.42, I
2
=0.0%). However, the pooled results
showed no signicant effect on IL-6 (WMD: 0.09 pg/ml, 95% CI: −0.22,
0.39, P =0.57) and brinogen (WMD: −6.07 mg/dL, 95% CI: −20.58,
8.44, P =0.41). Moreover, no evidence of heterogeneity was shown for
the RCTs evaluating IL-6 (Cochran Q test, P =0.401, I
2
=0.0%) and
brinogen (Cochran Q test, P =0.765, I
2
=0.0%) (Fig. 6).
3.3.6. Sensitivity analysis
The leave-one-out method was applied to evaluate the inuence of
each individual study on the pooled effect size (Kocaguneli & Menzies,
2013). The results remained robust after the sequential elimination of
RCTs for all outcomes (Supplementary Figure 7, 8, 9, 10, 11).
3.3.7. Publication bias
The visual inspection of the funnel plots and the Egger’s test revealed
no evidence of publication bias in the present study for weight (P =
0.25), BMI (P =0.59), WC (P =0.54), SBP (P =0.48), DBP (P =0.21),
a) b)
c)
NOTE: Weights are from random effects analysis
Overall (I-squared = 0.0%, p = 1.000)
ID
Study
Mohammad-shahi (2015)
Miraghajani (2013)
Faghih (2009)
Lukaszuk (2007)
Keshavarz (2012)
Azadbakht (2011)
Eslami (2018)
-0.74 (-1.65, 0.17)
WMD (95% CI)
-0.36 (-9.14, 8.42)
0.14 (-7.30, 7.58)
-0.82 (-1.88, 0.24)
-0.51 (-2.76, 1.74)
-0.16 (-6.09, 5.77)
-1.50 (-6.63, 3.63)
-0.15 (-6.01, 5.71)
100.00
Weight
%
1.07
1.50
73.26
16.26
2.35
3.14
2.41
-0.74 (-1.65, 0.17)
WMD (95% CI)
-0.36 (-9.14, 8.42)
0.14 (-7.30, 7.58)
-0.82 (-1.88, 0.24)
-0.51 (-2.76, 1.74)
-0.16 (-6.09, 5.77)
-1.50 (-6.63, 3.63)
-0.15 (-6.01, 5.71)
100.00
Weight
%
1.07
1.50
73.26
16.26
2.35
3.14
2.41
0-9.14 0 9.14
NOTE: Weights are from random effects analysis
Overall (I-squared = 0.0%, p = 1.000)
Mohammad-shahi (2015)
Azadbakht (2011)
ID
Faghih (2009)
Eslami (2018)
Keshavarz (2012)
Lukaszuk (2007)
Study
-0.21 (-0.51, 0.09)
-0.09 (-3.38, 3.20)
-0.20 (-1.74, 1.34)
WMD (95% CI)
-0.20 (-0.54, 0.14)
-0.04 (-1.80, 1.72)
-0.08 (-2.39, 2.23)
-0.32 (-1.14, 0.50)
100.00
0.84
3.84
Weight
77.28
2.92
1.71
13.41
%
-0.21 (-0.51, 0.09)
-0.09 (-3.38, 3.20)
-0.20 (-1.74, 1.34)
WMD (95% CI)
-0.20 (-0.54, 0.14)
-0.04 (-1.80, 1.72)
-0.08 (-2.39, 2.23)
-0.32 (-1.14, 0.50)
100.00
0.84
3.84
Weight
77.28
2.92
1.71
13.41
%
0-3.38 0 3.38
NOTE: Weights are from random effects analysis
Overall (I-squared = 0.0%, p = 0.847)
Eslami (2018)
Azadbakht (2011)
Keshavarz (2012)
Lukaszuk (2007)
Faghih (2009)
ID
Study
-1.61 (-2.74, -0.48)
-0.41 (-3.51, 2.69)
-0.10 (-4.51, 4.31)
-1.91 (-6.83, 3.01)
-2.62 (-6.92, 1.68)
-1.86 (-3.23, -0.49)
WMD (95% CI)
100.00
13.27
6.55
5.25
6.89
68.05
Weight
%
-1.61 (-2.74, -0.48)
-0.41 (-3.51, 2.69)
-0.10 (-4.51, 4.31)
-1.91 (-6.83, 3.01)
-2.62 (-6.92, 1.68)
-1.86 (-3.23, -0.49)
WMD (95% CI)
100.00
13.27
6.55
5.25
6.89
68.05
Weight
%
0-6.92 0 6.92
Fig. 4. Forest plots from the meta-analysis of clinical trials investigating the effects of soy milk supplementation on (a) body weight, (b) body mass index (BMI), c)
waist circumference (WC). WMD: weighted mean difference. CI: condence intervals.
a) b)
NOTE: Weights are from random effects analysis
Overall (I-squared = 89.9%, p = 0.000)
Azadbakht (2011)
Rivas (2002)
Maleki (2019)
Miraghajani (2013)
Study
Keshavarz (2012)
ID
-7.38 (-10.87, -3.88)
-2.70 (-8.53, 3.13)
-17.00 (-22.65, -11.35)
-2.33 (-4.12, -0.54)
-13.20 (-17.65, -8.75)
-5.00 (-5.58, -4.42)
WMD (95% CI)
100.00
15.20
15.59
24.62
18.46
%
26.13
Weight
-7.38 (-10.87, -3.88)
-2.70 (-8.53, 3.13)
-17.00 (-22.65, -11.35)
-2.33 (-4.12, -0.54)
-13.20 (-17.65, -8.75)
-5.00 (-5.58, -4.42)
WMD (95% CI)
100.00
15.20
15.59
24.62
18.46
%
26.13
Weight
0-22.7 0 22.7
NOTE: Weights are from random effects analysis
Overall (I-squared = 89.0%, p = 0.000)
Maleki (2019)
Study
ID
Miraghajani (2013)
Rivas (2002)
Keshavarz (2012)
Azadbakht (2011)
-4.36 (-7.06, -1.66)
-1.45 (-2.83, -0.07)
WMD (95% CI)
-8.10 (-10.98, -5.22)
-12.20 (-17.02, -7.38)
-1.70 (-2.28, -1.12)
-0.50 (-6.34, 5.34)
100.00
25.51
%
Weight
20.87
14.61
27.00
12.00
-4.36 (-7.06, -1.66)
-1.45 (-2.83, -0.07)
WMD (95% CI)
-8.10 (-10.98, -5.22)
-12.20 (-17.02, -7.38)
-1.70 (-2.28, -1.12)
-0.50 (-6.34, 5.34)
100.00
25.51
%
Weight
20.87
14.61
27.00
12.00
0-17 0 17
Fig. 5. Forest plots from the meta-analysis of clinical trials investigating the effects of soy milk supplementation on (a) systolic blood pressure (SBP) and (b) diastolic
blood pressure (DBP). WMD: weighted mean difference. CI: condence intervals.
M. Hassan Sohouli et al.

Journal of Functional Foods 83 (2021) 104499
10
TC (P =0.79), LDL-C (P =0.48), HDL-C (P =0.98), TG (P =0.80), FBG
(P =0.87), fasting insulin (P =0.23), TNF-
α
(P =0.91), IL-6 (P =0.68),
brinogen (P =0.17), and CRP (P =0.13) levels (Supplementary
Figure 12, 13, 14, 15, 16).
3.3.8. Discussion
In this systematic review and meta-analysis, we evaluated the effects
of the consumption of soy milk on various anthropometric indices and
cardiometabolic risk factors. We included all the relevant RCTs (n =18),
which included both healthy and subjects diagnosed with several car-
diometabolic diseases, to powerfully consolidate the conclusions and
overcome the contradictory and sample size limitations of individual
RCTs. Overall, the included RCTs had a low to unclear risk of bias and
comprised a total of 665 individuals (336 and 329 individuals received
soy milk and dairy milk/control diet, respectively). Our ndings
depicted benecial effects of soy milk consumption on blood pressure
and several serum lipids, inammation markers and anthropometric
indices. Specically, soy milk consumption reduced SBP, DBP and WC.
Moreover, it lowered TC and LDL-C. Lastly, soy milk administration
decreased CRP and TNF-
α
. On the other hand, soy milk consumption did
not exert any effects on FBG, fasting insulin, TG, several inammatory
markers (IL-6 and brinogen), and body composition indices (body
weight and BMI). Overall, our results reinforce that soy milk should be
integrated in the diet, owing to its favorable outcomes on blood pres-
sure, serum lipids, inammatory markers and anthropometric parame-
ters. Recent publications have also reported that soy products, for
example soy nuts or soy milk, exhibit positive effects on cardiovascular
risk factors. Whole soy, due to its composition in phytosterols, essential
fats, plant amino acids, and isoavones, might positively impact the
human health (Azadbakht & Nurbakhsh, 2011; Steinberg, 2007). Pure
phytoestrogens or isolated soy protein alone do not appear to be as
effective as combinations of soy with proteins, fatty acids, and phy-
toestrogens (Azadbakht et al., 2007). Several researchers have assessed
the effects of different soy products by comparing the pharmacokinetics
of isoavones from soymilk (liquid matrix) with those from textured
vegetable proteins (solid matrix). They discovered that, despite equiv-
alent doses per kilogram body weight, soymilk yielded higher maximal
plasma isoavone concentrations and total areas under the curve, which
implies that the matrix of food can inuence its effects on the human
health (Cassidy et al., 2006).
Cardiometabolic disorders, namely obesity, diabetes mellitus and
hypertension, are major concerns to the public healthcare worldwide
(“Global, regional, and national age-sex-specic mortality for 282 cau-
ses of death in 195 countries and territories, 1980–2017: a systematic
analysis for the Global Burden of Disease Study 2017,” 2018). They
commonly bundle together, being referred to as cardiometabolic mul-
timorbidity, and are associated with extensive rates of disability and
death (Di Angelantonio et al., 2015). An extensive body of literature
underscores the signicance of dietary patterns as crucial determinants
in the development of cardiometabolic disorders (Miranda et al., 2019;
D. Mozaffarian, 2016). Moreover, in addition to pharmacotherapy,
nutritional habits are considered key elements in both the management
and prevention of these disease (Miranda et al., 2019; D. Mozaffarian,
2016). In this context, the consumption of soy products has been re-
ported to exhibit positive effects on the human health, particularly in
patients with cardiovascular and metabolic disorders, including a
reduction of the cancer-related or cardiovascular-related risk of death
(Nachvak et al., 2019; Ramdath, Padhi, Sarfaraz, Renwick, & Duncan,
2017; Soltanipour, Hasandokht, Soleimani, Mahdavi-Roshan, & Jalali,
2019). It has been advocated that the consumption of whole soy, rather
than its selective components, is preferable as it is associated with more
health benets (Reinwald, Akabas, & Weaver, 2010). Distinctively, soy
a) b)
c) d)
NOTE: Weights are from random effects analysis
Overall (I-squared = 0.0%, p = 0.930)
Mohammad-shahi (2015)
ID
Eslami (2018)
Study
Nourieh (2012)
Miraghajani (2012)
-1.07 (-1.13, -1.01)
-1.07 (-1.13, -1.01)
WMD (95% CI)
-0.96 (-1.73, -0.19)
-0.37 (-2.68, 1.94)
-0.77 (-5.12, 3.58)
100.00
99.29
Weight
0.62
%
0.07
0.02
-1.07 (-1.13, -1.01)
-1.07 (-1.13, -1.01)
WMD (95% CI)
-0.96 (-1.73, -0.19)
-0.37 (-2.68, 1.94)
-0.77 (-5.12, 3.58)
100.00
99.29
Weight
0.62
%
0.07
0.02
0-5.12 0 5.12
NOTE: Weights are from random effects analysis
Overall (I-squared = 0.0%, p = 0.420)
Mohammad-shahi (2015)
Beavers (2009)
Study
ID
Miraghajani (2012)
-0.30 (-0.55, -0.06)
-0.30 (-0.57, -0.03)
-0.62 (-1.34, 0.10)
WMD (95% CI)
0.14 (-0.73, 1.01)
100.00
80.40
11.64
%
Weight
7.96
-0.30 (-0.55, -0.06)
-0.30 (-0.57, -0.03)
-0.62 (-1.34, 0.10)
WMD (95% CI)
0.14 (-0.73, 1.01)
100.00
80.40
11.64
%
Weight
7.96
0-1.34 0 1.34
NOTE: Weights are from random effects analysis
Overall (I-squared = 0.0%, p = 0.401)
Miraghajani (2012)
Beavers (2009)
Mohammad-shahi (2015)
Nourieh (2012)
ID
Study
0.09 (-0.22, 0.39)
0.17 (-0.55, 0.89)
-0.43 (-1.34, 0.48)
0.13 (-0.23, 0.49)
3.23 (-1.67, 8.13)
WMD (95% CI)
100.00
17.89
11.20
70.53
0.38
Weight
%
0.09 (-0.22, 0.39)
0.17 (-0.55, 0.89)
-0.43 (-1.34, 0.48)
0.13 (-0.23, 0.49)
3.23 (-1.67, 8.13)
WMD (95% CI)
100.00
17.89
11.20
70.53
0.38
Weight
%
0- 8.13 0 8.13
NOTE: Weights are from random effects analysis
Overall (I-squared = 0.0%, p = 0.765)
ID
Miraghajani (2012)
Keshavarz (2012)
Study
Maleki (2019)
-6.07 (-20.58, 8.44)
WMD (95% CI)
-15.76 (-47.59, 16.07)
-1.21 (-23.72, 21.30)
-6.09 (-29.73, 17.55)
100.00
Weight
20.78
41.54
%
37.67
-6.07 (-20.58, 8.44)
WMD (95% CI)
-15.76 (-47.59, 16.07)
-1.21 (-23.72, 21.30)
-6.09 (-29.73, 17.55)
100.00
Weight
20.78
41.54
%
37.67
0-47.6 0 47.6
Fig. 6. Forest plots from the meta-analysis of clinical trials investigating the effects of soy milk supplementation on (a) C-reactive protein (CRP), (b) tumour necrosis
factor
α
(TNF-
α
), c) interleukin 6 (IL-6) and d) brinogen. WMD: weighted mean difference. CI: condence intervals.
M. Hassan Sohouli et al.

Journal of Functional Foods 83 (2021) 104499
11
milk and whole soy are rather equivalent, as both virtually encompass
the same benecial components, for example essential fatty acids, soy
isoavones, inositols, healthy cholesterol, and phytosterols (Reinwald
et al., 2010).
Our pooled analysis demonstrated that subjects who received soy
milk experienced a decrease in their SBP and DBP. The blood pressure-
lowering effects of soy milk can be largely attributed to its composition
in isoavonoids, namely genistein and equol as a metabolite of daidzein,
both of which have exerted vasorelaxant and diuretic activities in in vivo
studies (Gimenez et al., 1997; Gim´
enez et al., 1998; Mishra et al., 2000).
Moreover, soy milk contains bioactive peptides that counteract the ac-
tivity of the angiotensin-converting enzyme (ACE) and stimulate the
activity of bradykinin, decreasing thus blood pressure (Maleki et al.,
2019). Based on our ndings, soy milk reduced SBP and DBP and we
may hypothesize that it can be combined with conventional antihyper-
tensive drugs in clinical practice in order to manage hypertension. It is
relevant to underscore that even a reduction of 2 mmHg in SBP and DBP
can decrease ischemic heart disease-related and stroke-related mortality
by 7% and 10%, respectively (Khalesi, Sun, Buys, & Jayasinghe, 2014;
Lewington, Clarke, Qizilbash, Peto, & Collins, 2003).
The greater effect of soy milk consumption on lowering blood pres-
sure during less than or equal to 4 weeks of intervention in this study
may be due to the effects of soy on microbiota composition. Evidence
has shown that soy in the long run causes an imbalance of intestinal
microbiota by reducing the relative abundance of benecial intestinal
bacteria as well as increasing the production of trimethylamine-N-oxide,
which may not be benecial to intestinal health(Ashaolu, 2020; Huang,
Krishnan, Pham, Yu, & Wang, 2016). Previous studies have shown an
association between microbiota imbalance and impaired blood pressure
regulation(Jama, Kaye, & Marques, 2019; Toral et al., 2019). It has also
been shown that the metabolic components of soy, such as genistein,
which are responsible for the antihypertensive effects, have a very short
half-life that loses its benecial effects in the long run(Rivas, Garay,
Escanero, Cia, et al., 2002). In addition, soy milk in the long run reduces
the absorption of micronutrients useful in lowering blood pressure, such
as magnesium (Rivas, Garay, Escanero, Cia, et al., 2002).
Soy proteins are approved by the United States Food and Drug
Administration (FDA) and other similar governing bodies from various
countries, particularly for their clinical utility in decreasing LDL-C levels
(Xiao, 2008). The molecular mechanisms and benecial cholesterol-
lowering effects of soy proteins are well-documented in the literature,
including evidence from several high-quality meta-analyses of RCTs
(Blanco Mejia et al., 2019; Reynolds et al., 2006). However, the
magnitude of LDL-C reduction remains a point of controversy ranging
from as low as 5% (Zhan & Ho, 2005) to as high as 13% (Anderson,
Johnstone, & Cook-Newell, 1995), yet none of the studies that have
reported these results have specically examined the efcacy of soy milk
versus dairy milk on various anthropometric parameters and car-
diometabolic risk factors. On one hand, our results demonstrated that
soy milk consumption lowered TC and LDL-C. On the other hand, the
impact of soy milk on HDL-C and TG levels was not signicant. An
accumulating body of evidence highlights that the lipid-lowering effects
of soy milk on LDL-C levels are more notable in subjects diagnosed with
hypercholesterolemia rather than individuals who do not suffer from
dyslipidemia (Anderson & Bush, 2011; Tokede, Onabanjo, Yansane,
Gaziano, & Djouss´
e, 2015; Zhan & Ho, 2005). This observation supports,
to a larger degree, the use of soy milk in selected individuals (for
example, dyslipidemic subjects) in order to attain maximum therapeutic
benets.
Inammation, marked by an increased expression of pro-
inammatory cytokines such as TNF-
α
(Zhang et al., 2009) and IL-6
(Volpato et al., 2001), is a key feature in the initiation and progres-
sion of cardiovascular diseases, in addition to its ability to forecast
mortality. Soy isoavones intrinsically harbor estrogen-like actions
which interfere with the generation of TNF-
α
(Ito et al., 2001) and IL-6
(Miyamoto et al., 1999). The anti-inammatory activity of soy
isoavones has been previously documented in various in vitro and in
vivo studies (Chacko et al., 2005; Sadeghalvad, Mohammadi-Motlagh,
Karaji, & Mostafaie, 2019). Our ndings demonstrated that individuals
who consume soy milk exhibit a signicant decrease in TNF-
α
and CRP
levels. Whether the magnitude of reduction is enough to cause a phys-
iological benet remains a point for further research. However, soy milk
consumption did not alter IL-6 concentrations. Soy milk and goat milk
contain oligosaccharides that reduce intestinal inammation and the
concentrations of pro-inammatory cytokines (Lara-Villoslada et al.,
2006; Tsangalis & Shah, 2004). In addition, several studies have shown
that soy isoavones may reduce CRP and pro-inammatory cytokine
concentrations by inhibiting the nuclear factor kappa B (NF-κB) as a
major regulator of pro-inammatory mediator synthesis (Fanti, Asmis,
Stephenson, Sawaya, & Franke, 2006; Mohammad-Shahi et al., 2016).
Soy isoavones ameliorate insulin sensitivity and glucose homeo-
stasis via several mechanisms, namely by decreasing the activity of the
intestinal alpha-glucosidase (Hanhineva et al., 2010) and the glucose
transporter type-4 (Ha et al., 2012). Nonetheless, in our study, while soy
milk supplementation favorably increased and decreased fasting insulin
and glucose levels, respectively, these effects were not statistically sig-
nicant. These ndings were in agreement with the results of two meta-
analyses which revealed that soy consumption did not lead to clinically
meaningful effects on various indices of glycemic control in healthy
individuals or in patients diagnosed with cardiometabolic disorders, for
example diabetes mellitus (Liu, Chen, & Ho, 2011; Soltanipour et al.,
2019). Further research is needed to explore in which subjects soy milk
may show increased benets and whether the administration of this
product should be tailored according to certain clinical features or de-
mographics, for example glycemic status.
With regard to anthropometric parameters, our ndings revealed a
signicant reduction in WC following soy milk consumption. However,
in the current meta-analysis, no signicant changes were observed for
other anthropometric indices, such as body weight or BMI. A recent
meta-analysis revealed that the administration of soy foods and iso-
avones did not alter weight and WC in a signicant fashion (Akhlaghi,
Zare, & Nouripour, 2017). However, the subgroup analyses underpinned
that this effect could vary based on the age, the gender of the partici-
pants, the consumption dose, the type of soy products employed, the
duration of the intervention, as well as the anthropometric indices at
baseline (Akhlaghi et al., 2017). Therefore, such differences might
explain the conicting inter-study results regarding the impact of soy on
anthropometric indices. On the other hand, a meta-analysis of 43 RCTs
that examined the effects of soy protein and soy isoavones on anthro-
pometric parameters detected no statistically signicant changes of
these variables following the intervention (Mu et al., 2019). Various
studies with high sample sizes have shown that large waist sizes in men
and women are signicantly associated with low HDL-C and high levels
of fasting triacylglycerol, insulin, FBG, LDL-C, as well as elevated blood
pressure, all of which are markers of cardiovascular risk (Dobbelsteyn,
Joffres, MacLean, & Flowerdew, 2001; Seidell, P´
erusse, Despr´
es, &
Bouchard, 2001). Increased WC reects an elevated accumulation of
visceral fat and an overexposure of the liver to fatty acids, and can be
involved in the development of insulin resistance and endothelial
dysfunction, as well as inuence the lipid prole (Gans, 2006).
Our study has several strengths. To the best of our knowledge, this is
the rst systematic review and meta-analysis that scrutinized the impact
of soy milk consumption specically, as opposed to other soy products,
on cardiometabolic risk factors. The design of our study, that is sys-
tematic review and meta-analysis, is advantageous and enable us to draw
solid conclusions and to overcome the contradictions as well as sample
size limitations of individual RCTs. We included only papers with high-
quality study designs (RCTs) and excluded non-RCTs papers. We
analyzed a comprehensive panel of biochemical markers and anthro-
pometric indices, namely blood pressure and glycemic, lipid, inam-
matory, and anthropometric indices. The meta-analysis of the RCTs
evaluating changes in the SBP and DBP was marked by heterogeneity
M. Hassan Sohouli et al.

Journal of Functional Foods 83 (2021) 104499
12
and thus we attempted to resolve this drawback by employing subgroup
analyses based on the duration of the intervention and the health status
of the participants. The other assessed variables were homogeneous
(except for the lipid prole), reecting no substantial inter-study vari-
ations of the outcomes. Moreover, we employed leave-one-out sensi-
tivity analyses which reinforced the robustness of the ndings and we
did not identify publication biases in our study outcomes.
Nonetheless, our study has some limitations that ought to be recog-
nized. Notably, the small number of included RCTs and their sample
sizes constitute a major limitation. The included studies mostly recruited
women, hence the results might not be generalizable to men. Moreover,
another limitation of our paper is that, due to the small number of
studies in each group in terms of study populations, it was not possible to
perform subgroup analysis based on the health status of the participants,
which could have affected the interpretation of our results. In addition,
we discovered that the quality of most of the analyzed RCTs was het-
erogeneous. Furthermore, the adherence/compliance to the soy milk
intervention was not quantitatively veried by examining the serum or
urinary levels of soy isoavones.
3.3.9. Conclusions
According to the available evidence, soy milk consumption can exert
favorable effects on blood pressure, several components of the lipid
prole and certain anthropometric and inammatory markers as
compared to controls. However, further research on larger samples is
warranted to conrm these ndings.
Funding
None
Declaration of Competing Interest
The authors declare that they have no known competing nancial
interests or personal relationships that could have appeared to inuence
the work reported in this paper.
Appendix A. Supplementary data
Supplementary data to this article can be found online at https://doi.
org/10.1016/j.jff.2021.104499.
References
Akhlaghi, M., Zare, M., & Nouripour, F. (2017). Effect of Soy and Soy Isoavones on
Obesity-Related Anthropometric Measures: A Systematic Review and Meta-analysis
of Randomized Controlled Clinical Trials. Adv Nutr, 8(5), 705–717. https://doi.org/
10.3945/an.117.015370.
Anderson, J. W., & Bush, H. M. (2011). Soy protein effects on serum lipoproteins: A
quality assessment and meta-analysis of randomized, controlled studies. J Am Coll
Nutr, 30(2), 79–91. https://doi.org/10.1080/07315724.2011.10719947.
Anderson, J. W., Johnstone, B. M., & Cook-Newell, M. E. (1995). Meta-analysis of the
effects of soy protein intake on serum lipids. N Engl J Med, 333(5), 276–282. https://
doi.org/10.1056/nejm199508033330502.
Ashaolu, T. J. (2020). Soy bioactive peptides and the gut microbiota modulation. Applied
Microbiology and Biotechnology, 1–9.
Azadbakht, L., Kimiagar, M., Mehrabi, Y., Esmaillzadeh, A., Hu, F. B., & Willett, W. C.
(2007). Soy consumption, markers of inammation, and endothelial function: A
cross-over study in postmenopausal women with the metabolic syndrome. Diabetes
care, 30(4), 967–973.
Azadbakht, L., & Nurbakhsh, S. (2011). Effect of soy drink replacement in a weight
reducing diet on anthropometric values and blood pressure among overweight and
obese female youths. Asia Pacic journal of clinical nutrition, 20(3), 383.
Beavers, K. M., Serra, M. C., Beavers, D. P., Cooke, M. B., & Willoughby, D. S. (2009).
Soymilk supplementation does not alter plasma markers of inammation and
oxidative stress in postmenopausal women. Nutrition research, 29(9), 616–622.
Blanco Mejia, S., Messina, M., Li, S. S., Viguiliouk, E., Chiavaroli, L., Khan, T. A., & Kris-
Etherton, P. (2019). A meta-analysis of 46 studies identied by the FDA
demonstrates that soy protein decreases circulating LDL and total cholesterol
concentrations in adults. The Journal of nutrition, 149(6), 968–981.
Cassidy, A., Brown, J. E., Hawdon, A., Faughnan, M. S., King, L. J., Millward, J., &
Setchell, K. D. (2006). Factors affecting the bioavailability of soy isoavones in
humans after ingestion of physiologically relevant levels from different soy foods.
The Journal of nutrition, 136(1), 45–51.
Chacko, B. K., Chandler, R. T., Mundhekar, A., Khoo, N., Pruitt, H. M., Kucik, D. F., &
Patel, R. P. (2005). Revealing anti-inammatory mechanisms of soy isoavones by
ow: Modulation of leukocyte-endothelial cell interactions. Am J Physiol Heart Circ
Physiol, 289(2), H908–H915. https://doi.org/10.1152/ajpheart.00781.2004.
Danaei, G., Lu, Y., Singh, G. M., Carnahan, E., Stevens, G. A., Cowan, M. J., & Rao, M.
(2014). Cardiovascular disease, chronic kidney disease, and diabetes mortality
burden of cardiometabolic risk factors from 1980 to 2010: A comparative risk
assessment. Lancet Diabetes & Endocrinology.
Cunha, C.d. M., Costa, P. R. F., de Oliveira, L. P. M., Queiroz, V. A.d. O.,
Pitangueira, J. C. D., & Oliveira, A. M. (2018). Dietary patterns and cardiometabolic
risk factors among adolescents: Systematic review and meta-analysis. British Journal
of Nutrition, 119(8), 859–879.
Di Angelantonio, E., Kaptoge, S., Wormser, D., Willeit, P., Butterworth, A. S., Bansal, N.,
& Danesh, J. (2015). Association of Cardiometabolic Multimorbidity With Mortality.
Jama, 314(1), 52. https://doi.org/10.1001/jama.2015.7008.
Dobbelsteyn, C., Joffres, M., MacLean, D. R., & Flowerdew, G. (2001). A comparative
evaluation of waist circumference, waist-to-hip ratio and body mass index as
indicators of cardiovascular risk factors. The Canadian Heart Health Surveys.
International journal of obesity, 25(5), 652–661.
Egger, M., Smith, G. D., Schneider, M., & Minder, C. (1997). Bias in meta-analysis
detected by a simple, graphical test. Bmj, 315(7109), 629–634.
Eslami, O., Shidfar, F., Maleki, Z., Jazayeri, S., Hosseini, A. F., Agah, S., & Ardiyani, F.
(2019). Effect of soy milk on metabolic status of patients with nonalcoholic fatty
liver disease: A randomized clinical trial. Journal of the American College of Nutrition,
38(1), 51–58.
Faghih, S.h., Abadi, A. R., Hedayati, M., & Kimiagar, S. M. (2011). Comparison of the
effects of cows’ milk, fortied soy milk, and calcium supplement on weight and fat
loss in premenopausal overweight and obese women. Nutrition, Metabolism and
Cardiovascular Diseases, 21(7), 499–503.
Faghih, S., Hedayati, M., Abadi, A., & Kimiagar, M. (2009). Comparison of the Effects of
Cow’ s Milk, Fortied Soy Milk, and Calcium Supplement on Plasma Adipocytokines
in Overweight and Obese Women. Iranian Journal of Endocrinology and
Metabolism, 11(6), 692-698.
Fanti, P., Asmis, R., Stephenson, T. J., Sawaya, B. P., & Franke, A. A. (2006). Positive
effect of dietary soy in ESRD patients with systemic inammation—correlation
between blood levels of the soy isoavones and the acute-phase reactants.
Nephrology Dialysis Transplantation, 21(8), 2239–2246.
Gans, R. O. B. (2006). The metabolic syndrome, depression, and cardiovascular disease:
Interrelated conditions that share pathophysiologic mechanisms. Medical Clinics, 90
(4), 573–591.
Gardner, C. D., Messina, M., Kiazand, A., Morris, J. L., & Franke, A. A. (2007). Effect of
two types of soy milk and dairy milk on plasma lipids in hypercholesterolemic
adults: A randomized trial. Journal of the American College of Nutrition, 26(6),
669–677.
Gimenez, I., Lou, M., Vargas, F., Alvarez-Guerra, M., Mayoral, J. A., Martinez, R. M., &
Alda, J. O. (1997). Renal and vascular actions of equol in the rat. J Hypertens, 15(11),
1303–1308. https://doi.org/10.1097/00004872-199715110-00015.
Gim´
enez, I., Martinez, R. M., Lou, M., Mayoral, J. A., Garay, R. P., & Alda, J. O. (1998).
Salidiuretic action by genistein in the isolated, perfused rat kidney. Hypertension, 31
(2), 706–711. https://doi.org/10.1161/01.hyp.31.2.706.
Ha, B. G., Nagaoka, M., Yonezawa, T., Tanabe, R., Woo, J. T., Kato, H., & Yagasaki, K.
(2012). Regulatory mechanism for the stimulatory action of genistein on glucose
uptake in vitro and in vivo. J Nutr Biochem, 23(5), 501–509. https://doi.org/
10.1016/j.jnutbio.2011.02.007.
Hanhineva, K., T¨
orr¨
onen, R., Bondia-Pons, I., Pekkinen, J., Kolehmainen, M.,
Mykk¨
anen, H., & Poutanen, K. (2010). Impact of dietary polyphenols on
carbohydrate metabolism. Int J Mol Sci, 11(4), 1365–1402. https://doi.org/10.3390/
ijms11041365.
Higgins, J. (2011). Cochrane handbook for systematic reviews of interventions. Version
5.1. 0 [updated March 2011]. The Cochrane Collaboration. www. cochrane-
handbook. org.
Higgins, J., & Wells, G. (2011). Cochrane handbook for systematic reviews of
interventions.
Higgins, J. P., Thompson, S. G., Deeks, J. J., & Altman, D. G. (2003). Measuring
inconsistency in meta-analyses. Bmj, 327(7414), 557–560.
Hozo, S. P., Djulbegovic, B., & Hozo, I. (2005). Estimating the mean and variance from
the median, range, and the size of a sample. BMC medical research methodology, 5(1),
13.
Huang, H., Krishnan, H. B., Pham, Q., Yu, L. L., & Wang, T. T. Y. (2016). Soy and gut
microbiota: Interaction and implication for human health. Journal of agricultural and
food chemistry, 64(46), 8695–8709.
Ito, A., Bebo, B. F., Jr., Matejuk, A., Zamora, A., Silverman, M., Fyfe-Johnson, A., &
Offner, H. (2001). Estrogen treatment down-regulates TNF-alpha production and
reduces the severity of experimental autoimmune encephalomyelitis in cytokine
knockout mice. J Immunol, 167(1), 542–552. https://doi.org/10.4049/
jimmunol.167.1.542.
Jama, H. A., Kaye, D. M., & Marques, F. Z. (2019). The gut microbiota and blood pressure
in experimental models. Current opinion in nephrology and hypertension, 28(2),
97–104.
Keshavarz, S. A., Nourieh, Z., Attar, M. J. H., & Azadbakht, L. (2012). Effect of soymilk
consumption on waist circumference and cardiovascular risks among overweight
and obese female adults. International journal of preventive medicine, 3(11), 798.
M. Hassan Sohouli et al.

Journal of Functional Foods 83 (2021) 104499
13
Khalesi, S., Sun, J., Buys, N., & Jayasinghe, R. (2014). Effect of probiotics on blood
pressure: A systematic review and meta-analysis of randomized, controlled trials.
Hypertension, 64(4), 897–903.
Kocaguneli, E., & Menzies, T. (2013). Software effort models should be assessed via
leave-one-out validation. Journal of Systems and Software, 86(7), 1879–1890.
Lara-Villoslada, F., Debras, E., Nieto, A., Concha, A., G´
alvez, J., L´
opez-Huertas, E., &
Xaus, J. (2006). Oligosaccharides isolated from goat milk reduce intestinal
inammation in a rat model of dextran sodium sulfate-induced colitis. Clinical
Nutrition, 25(3), 477–488.
Lewington, S., Clarke, R., Qizilbash, N., Peto, R., & Collins, R. (2003). Age-specic
relevance of usual blood pressure to vascular mortality. The Lancet, 361(9366),
1391–1392.
Liu, Z. M., Chen, Y. M., & Ho, S. C. (2011). Effects of soy intake on glycemic control: A
meta-analysis of randomized controlled trials. Am J Clin Nutr, 93(5), 1092–1101.
https://doi.org/10.3945/ajcn.110.007187.
Lukaszuk, J. M., Luebbers, P., & Gordon, B. A. (2007). Preliminary study: Soy milk as
effective as skim milk in promoting weight loss. Journal of the American Dietetic
Association, 107(10), 1811–1814.
Maleki, Z., Jazayeri, S., Eslami, O., Shidfar, F., Hosseini, A. F., Agah, S., & Norouzi, H.
(2019). Effect of soy milk consumption on glycemic status, blood pressure,
brinogen and malondialdehyde in patients with non-alcoholic fatty liver disease: A
randomized controlled trial. Complementary Therapies in Medicine, 44, 44–50.
Mazumder, M. A. R., & Begum, A. A. (2016). Soy milk as source of nutrient for
malnourished population of developing country: A review. International Journal of
Advanced Scientic and Technical Research, 5(6), 192–203.
McQueen, R. B., Ghushchyan, V., Olufade, T., Sheehan, J. J., Nair, K. V., & Saseen, J. J.
(2016). Incremental increases in economic burden parallels cardiometabolic risk
factors in the US. Diabetes, Metabolic Syndrome and Obesity: Targets and Therapy, 9,
233.
Micha, R., Pe˜
nalvo, J. L., Cudhea, F., Imamura, F., Rehm, C. D., & Mozaffarian, D. (2017).
Association between dietary factors and mortality from heart disease, stroke, and
type 2 diabetes in the United States. Jama, 317(9), 912–924.
Miraghajani, M. S., Esmaillzadeh, A., Najafabadi, M. M., Mirlohi, M., & Azadbakht, L.
(2012). Soy milk consumption, inammation, coagulation, and oxidative stress
among type 2 diabetic patients with nephropathy. Diabetes care, 35(10), 1981–1985.
Miraghajani, M. S., Najafabadi, M. M., Surkan, P. J., Esmaillzadeh, A., Mirlohi, M., &
Azadbakht, L. (2013). Soy milk consumption and blood pressure among type 2
diabetic patients with nephropathy. Journal of Renal Nutrition, 23(4), 277–282.e1.
Miranda, J. J., Barrientos-Guti´
errez, T., Corvalan, C., Hyder, A. A., Lazo-Porras, M.,
Oni, T., & Wells, J. C. K. (2019). Understanding the rise of cardiometabolic diseases
in low- and middle-income countries. Nat Med, 25(11), 1667–1679. https://doi.org/
10.1038/s41591-019-0644-7.
Mishra, S. K., Abbot, S. E., Choudhury, Z., Cheng, M., Khatab, N., Maycock, N. J., &
Aaronson, P. I. (2000). Endothelium-dependent relaxation of rat aorta and main
pulmonary artery by the phytoestrogens genistein and daidzein. Cardiovasc Res, 46
(3), 539–546. https://doi.org/10.1016/s0008-6363(00)00049-3.
Mitchell, J. H., & Collins, A. R. (1999). Effects of a soy milk supplement on plasma
cholesterol levels and oxidative DNA damage in men–a pilot study. European Journal
of Nutrition, 38(3), 143–1138.
Miyamoto, N., Mandai, M., Suzuma, I., Suzuma, K., Kobayashi, K., & Honda, Y. (1999).
Estrogen protects against cellular inltration by reducing the expressions of E-
selectin and IL-6 in endotoxin-induced uveitis. J Immunol, 163(1), 374–379.
Mohammad-Shahi, M., Mowla, K., Haidari, F., Zarei, M., & Choghakhori, R. (2016). Soy
milk consumption, markers of inammation and oxidative stress in women with
rheumatoid arthritis: A randomised cross-over clinical trial. Nutrition & Dietetics, 73
(2), 139–145.
Moher, D., Shamseer, L., Clarke, M., Ghersi, D., Liberati, A., Petticrew, M., &
Stewart, L. A. (2015). Preferred reporting items for systematic review and meta-
analysis protocols (PRISMA-P) 2015 statement. Systematic reviews, 4(1). https://doi.
org/10.1186/2046-4053-4-1.
Mozaffarian, D. (2016). Dietary and Policy Priorities for Cardiovascular Disease,
Diabetes, and Obesity: A Comprehensive Review. Circulation, 133(2), 187–225.
https://doi.org/10.1161/circulationaha.115.018585.
Mozaffarian, D., Wilson, P. W. F., & Kannel, W. B. (2008). Beyond established and novel
risk factors: Lifestyle risk factors for cardiovascular disease. Circulation, 117(23),
3031–3038.
Mu, Y., Kou, T., Wei, B., Lu, X., Liu, J., Tian, H., & Cui, W. (2019). Soy products
ameliorate obesity-related anthropometric indicators in overweight or obese asian
and non-menopausal women: A meta-analysis of randomized controlled trials.
Nutrients, 11(11), 2790.
Nachvak, S. M., Moradi, S., Anjom-Shoae, J., Rahmani, J., Nasiri, M., Maleki, V., &
Sadeghi, O. (2019). Soy, Soy Isoavones, and Protein Intake in Relation to Mortality
from All Causes, Cancers, and Cardiovascular Diseases: A Systematic Review and
Dose-Response Meta-Analysis of Prospective Cohort Studies. J Acad Nutr Diet, 119
(9), 1483–1500.e1417. https://doi.org/10.1016/j.jand.2019.04.011.
Nourieh, Z., Keshavarz, S., HosseinzadehAttar, M., & Azadbakht, L. (2012). Effect of
soymilk consumption on Inammatory markers and lipid proles among non-
menopausal overweight and obese female adults. Journal of Research in Medical
Sciences, 17(4), 1.
O’Keefe, J. H., Gheewala, N. M., & O’Keefe, J. O. (2008). Dietary strategies for improving
post-prandial glucose, lipids, inammation, and cardiovascular health. Journal of the
American College of Cardiology, 51(3), 249–255.
¨
Onning, G., Åkesson, B., ¨
Oste, R., & Lundquist, I. (1998). Effects of consumption of oat
milk, soya milk, or cow’s milk on plasma lipids and antioxidative capacity in healthy
subjects. Annals of nutrition and metabolism, 42(4), 211–220.
Onuegbu, A. J., Olisekodiaka, J. M., Onibon, M. O., Adesiyan, A. A., & Igbeneghu, C. A.
(2011). Consumption of soymilk lowers atherogenic lipid fraction in healthy
individuals. Journal of medicinal food, 14(3), 257–260.
Ramdath, D., Padhi, E., Sarfaraz, S., Renwick, S., & Duncan, A. (2017). Beyond the
Cholesterol-Lowering Effect of Soy Protein: A Review of the Effects of Dietary Soy
and Its Constituents on Risk Factors for Cardiovascular Disease. Nutrients, 9(4), 324.
https://doi.org/10.3390/nu9040324.
Rao, G. H. (2018). Prevention or reversal of cardiometabolic diseases. Journal of Clinical
and Preventive Cardiology, 7(1), 22. https://doi.org/10.4103/JCPC.JCPC_41_17.
Reinwald, S., Akabas, S. R., & Weaver, C. M. (2010). Whole versus the piecemeal
approach to evaluating soy. J Nutr, 140(12), 2335s–2343s. https://doi.org/10.3945/
jn.110.124925.
Reynolds, K., Chin, A., Lees, K. A., Nguyen, A., Bujnowski, D., & He, J. (2006). A meta-
analysis of the effect of soy protein supplementation on serum lipids. The American
journal of cardiology, 98(5), 633–640.
Rivas, M., Garay, R. P., Escanero, J. s. F., Cia Jr, P., Cia, P., & Alda, J. O. (2002). Soy milk
lowers blood pressure in men and women with mild to moderate essential
hypertension. The Journal of nutrition, 132(7), 1900-1902.
Sadeghalvad, M., Mohammadi-Motlagh, H. R., Karaji, A. G., & Mostafaie, A. (2019). In
vivo anti-inammatory efcacy of the combined Bowman-Birk trypsin inhibitor and
genistein isoavone, two biological compounds from soybean. Journal of
biochemical and molecular toxicology, 33(12), e22406.
Seidell, J. C., P´
erusse, L., Despr´
es, J.-P., & Bouchard, C. (2001). Waist and hip
circumferences have independent and opposite effects on cardiovascular disease risk
factors: The Quebec Family Study. The American Journal of Clinical Nutrition, 74(3),
315–321. https://doi.org/10.1093/ajcn/74.3.315.
Sethi, S., Tyagi, S. K., & Anurag, R. K. (2016). Plant-based milk alternatives an emerging
segment of functional beverages: A review. Journal of food science and technology, 53
(9), 3408–3423.
Soltanipour, S., Hasandokht, T., Soleimani, R., Mahdavi-Roshan, M., & Jalali, M. M.
(2019). Systematic Review and Meta-Analysis of the Effects of Soy on Glucose
Metabolism in Patients with Type 2 Diabetes. Rev Diabet Stud, 15(1), 60–70. https://
doi.org/10.1900/RDS.2019.15.60.
Steinberg, F. M. (2007). Soybeans or soymilk: does it make a difference for
cardiovascular protection? Does it even matter? : Oxford University Press.
Takatsuka, N., Nagata, C., Kurisu, Y., Inaba, S., Kawakami, N., & Shimizu, H. (2000).
Hypocholesterolemic Effect of Soymilk Supplementation with Usual Diet in
Premenopausal Normolipidemic Japanese Women. Preventive Medicine, 31(4),
308–314. https://doi.org/10.1006/pmed.2000.0714.
Tokede, O. A., Onabanjo, T. A., Yansane, A., Gaziano, J. M., & Djouss´
e, L. (2015). Soya
products and serum lipids: A meta-analysis of randomised controlled trials. Br J Nutr,
114(6), 831–843. https://doi.org/10.1017/s0007114515002603.
Toral, M., Robles-Vera, I., de la Visitaci´
on, N., Romero, M., Yang, T., S´
anchez, M., &
Duarte, J. (2019). Critical role of the interaction gut microbiota–sympathetic
nervous system in the regulation of blood pressure. Frontiers in physiology, 10.
https://doi.org/10.3389/fphys.2019.00231.
Tsangalis, D., & Shah, N. P. (2004). Metabolism of oligosaccharides and aldehydes and
production of organic acids in soymilk by probiotic bidobacteria. International
journal of food science & technology, 39(5), 541–554.
Volpato, S., Guralnik, J. M., Ferrucci, L., Balfour, J., Chaves, P., Fried, L. P., &
Harris, T. B. (2001). Cardiovascular disease, interleukin-6, and risk of mortality in
older women: The women’s health and aging study. Circulation, 103(7), 947–953.
https://doi.org/10.1161/01.cir.103.7.947.
Wang, J., Wu, R., Zhang, W., Sun, Z., Zhao, W., & Zhang, H. (2013). Proteomic
comparison of the probiotic bacterium Lactobacillus casei Zhang cultivated in milk
and soy milk. Journal of dairy science, 96(9), 5603–5624.
Woodside, J. V., Brennan, S., & Cantwell, M. (2016). Are Soy-Milk Products Viable
Alternatives to Cow’s Milk? Beverage Impacts on Health and Nutrition (pp. 151-
162): Springer.
Xiao, C. W. (2008). Health effects of soy protein and isoavones in humans. J Nutr, 138
(6), 1244s–1249s. https://doi.org/10.1093/jn/138.6.1244S.
Zhan, S., & Ho, S. C. (2005). Meta-analysis of the effects of soy protein containing
isoavones on the lipid prole. Am J Clin Nutr, 81(2), 397–408. https://doi.org/
10.1093/ajcn.81.2.397.
Zhang, H., Park, Y., Wu, J., Chen, X., Lee, S., Yang, J., & Zhang, C. (2009). Role of TNF-
alpha in vascular dysfunction. Clin Sci (Lond), 116(3), 219–230. https://doi.org/
10.1042/cs20080196.
M. Hassan Sohouli et al.