Skip to main content


  • Access
  • Cited by 35
  • Cited by
    This article has been cited by the following publications. This list is generated based on data provided by CrossRef.

    Kohno, Mitsutaka Sugano, Hideo Shigihara, Yuhko Shiraishi, Yoshiaki and Motoyama, Takayasu 2018. Improvement of glucose and lipid metabolism via mung bean protein consumption: clinical trials of GLUCODIA™ isolated mung bean protein in the USA and Canada. Journal of Nutritional Science, Vol. 7, Issue. ,

    Dreher, Mark L. 2018. Dietary Patterns and Whole Plant Foods in Aging and Disease. p. 337.

    Afiat, Maliheh Amini, Elham Ghazanfarpour, Masumeh Nouri, Bahareh Mousavi, Mahsa Sadat Babakhanian, Masoudeh and Rakhshandeh, Hassan 2018. The Effect of Short-term Treatment with Fennel on Lipid Profile in Postmenopausal Women: A Randomized Controlled Trial. Journal of Menopausal Medicine, Vol. 24, Issue. 1, p. 29.

    Burkard, Marion Huth, Karl and Leitzmann, Claus 2018. Ernährung und Fasten als Therapie. p. 227.

    Ristić-Medić, Danijela Takić, Marija and Radjen, Slavica 2018. Therapeutic Foods. p. 329.

    Banach, Maciej Patti, Angelo Maria Giglio, Rosaria Vincenza Cicero, Arrigo F.G. Atanasov, Atanas G. Bajraktari, Gani Bruckert, Eric Descamps, Olivier Djuric, Dragan M. Ezhov, Marat Fras, Zlatko von Haehling, Stephan Katsiki, Niki Langlois, Michel Latkovskis, Gustavs Mancini, G.B. John Mikhailidis, Dimitri P. Mitchenko, Olena Moriarty, Patrick M. Muntner, Paul Nikolic, Dragana Panagiotakos, Demosthenes B. Paragh, Gyorgy Paulweber, Bernhard Pella, Daniel Pitsavos, Christos Reiner, Željko Rosano, Giuseppe M.C. Rosenson, Robert S. Rysz, Jacek Sahebkar, Amirhossein Serban, Maria-Corina Vinereanu, Dragos Vrablík, Michal Watts, Gerald F. Wong, Nathan D. and Rizzo, Manfredi 2018. The Role of Nutraceuticals in Statin Intolerant Patients. Journal of the American College of Cardiology, Vol. 72, Issue. 1, p. 96.

    Bianconi, Vanessa Mannarino, Massimo Raffaele Sahebkar, Amirhossein Cosentino, Teresa and Pirro, Matteo 2018. Cholesterol-Lowering Nutraceuticals Affecting Vascular Function and Cardiovascular Disease Risk. Current Cardiology Reports, Vol. 20, Issue. 7,

    Cicero, Arrigo F. G. and Colletti, Alessandro 2018. An update on the safety of nutraceuticals and effects on lipid parameters. Expert Opinion on Drug Safety, Vol. 17, Issue. 3, p. 303.

    Rizzo, Gianluca and Baroni, Luciana 2018. Soy, Soy Foods and Their Role in Vegetarian Diets. Nutrients, Vol. 10, Issue. 1, p. 43.

    Poli, Andrea Barbagallo, Carlo M. Cicero, Arrigo F.G. Corsini, Alberto Manzato, Enzo Trimarco, Bruno Bernini, Franco Visioli, Francesco Bianchi, Alfio Canzone, Giuseppe Crescini, Claudio de Kreutzenberg, Saula Ferrara, Nicola Gambacciani, Marco Ghiselli, Andrea Lubrano, Carla Marelli, Giuseppe Marrocco, Walter Montemurro, Vincenzo Parretti, Damiano Pedretti, Roberto Perticone, Francesco Stella, Roberto and Marangoni, Franca 2018. Nutraceuticals and functional foods for the control of plasma cholesterol levels. An intersociety position paper. Pharmacological Research, Vol. 134, Issue. , p. 51.

    Sievenpiper, John L. Chan, Catherine B. Dworatzek, Paula D. Freeze, Catherine and Williams, Sandra L. 2018. Nutrition Therapy. Canadian Journal of Diabetes, Vol. 42, Issue. , p. S64.

    Jelena, Cvejić Hogervorst Giorgio, Russo Justyna, Godos Neda, Mimica-Dukić Natasa, Simin Artur, Bjelica and Giuseppe, Grosso 2018. Polyphenols: Properties, Recovery, and Applications. p. 69.

    Simental-Mendía, Luis E. Gotto, Antonio M. Atkin, Stephen L. Banach, Maciej Pirro, Matteo and Sahebkar, Amirhossein 2018. Effect of soy isoflavone supplementation on plasma lipoprotein(a) concentrations: A meta-analysis. Journal of Clinical Lipidology, Vol. 12, Issue. 1, p. 16.

    García-Conesa, María-Teresa Chambers, Karen Combet, Emilie Pinto, Paula Garcia-Aloy, Mar Andrés-Lacueva, Cristina de Pascual-Teresa, Sonia Mena, Pedro Konic Ristic, Aleksandra Hollands, Wendy Kroon, Paul Rodríguez-Mateos, Ana Istas, Geoffrey Kontogiorgis, Christos Rai, Dilip Gibney, Eileen Morand, Christine Espín, Juan and González-Sarrías, Antonio 2018. Meta-Analysis of the Effects of Foods and Derived Products Containing Ellagitannins and Anthocyanins on Cardiometabolic Biomarkers: Analysis of Factors Influencing Variability of the Individual Responses. International Journal of Molecular Sciences, Vol. 19, Issue. 3, p. 694.

    Zou, Yuan Pan, Runting Ruan, Qijun Wan, Zhili Guo, Jian and Yang, Xiaoquan 2018. Interaction of Soybean 7S Globulin Peptide with Cell Membrane Model via Isothermal Titration Calorimetry, Quartz Crystal Microbalance with Dissipation, and Langmuir Monolayer Study. Journal of Agricultural and Food Chemistry, Vol. 66, Issue. 19, p. 4913.

    Orona-Tamayo, Domancar Valverde, María Elena and Paredes-López, Octavio 2018. Bioactive peptides from selected latin american food crops – A nutraceutical and molecular approach. Critical Reviews in Food Science and Nutrition, p. 1.

    Karamali, M. Kashanian, M. Alaeinasab, S. and Asemi, Z. 2018. The effect of dietary soy intake on weight loss, glycaemic control, lipid profiles and biomarkers of inflammation and oxidative stress in women with polycystic ovary syndrome: a randomised clinical trial. Journal of Human Nutrition and Dietetics, Vol. 31, Issue. 4, p. 533.

    Dreher, Mark L. 2018. Dietary Patterns and Whole Plant Foods in Aging and Disease. p. 59.

    Namazi, Nazli Saneei, Parvane Larijani, Bagher and Esmaillzadeh, Ahmad 2018. Soy product consumption and the risk of all-cause, cardiovascular and cancer mortality: a systematic review and meta-analysis of cohort studies. Food & Function, Vol. 9, Issue. 5, p. 2576.

    Cicero, Arrigo F G Fogacci, Federica and Colletti, Alessandro 2017. Potential role of bioactive peptides in prevention and treatment of chronic diseases: a narrative review. British Journal of Pharmacology, Vol. 174, Issue. 11, p. 1378.



      • Send article to Kindle

        To send this article to your Kindle, first ensure is added to your Approved Personal Document E-mail List under your Personal Document Settings on the Manage Your Content and Devices page of your Amazon account. Then enter the ‘name’ part of your Kindle email address below. Find out more about sending to your Kindle. Find out more about sending to your Kindle.

        Note you can select to send to either the or variations. ‘’ emails are free but can only be sent to your device when it is connected to wi-fi. ‘’ emails can be delivered even when you are not connected to wi-fi, but note that service fees apply.

        Find out more about the Kindle Personal Document Service.

        Soya products and serum lipids: a meta-analysis of randomised controlled trials
        Available formats
        Send article to Dropbox

        To send this article to your Dropbox account, please select one or more formats and confirm that you agree to abide by our usage policies. If this is the first time you use this feature, you will be asked to authorise Cambridge Core to connect with your <service> account. Find out more about sending content to Dropbox.

        Soya products and serum lipids: a meta-analysis of randomised controlled trials
        Available formats
        Send article to Google Drive

        To send this article to your Google Drive account, please select one or more formats and confirm that you agree to abide by our usage policies. If this is the first time you use this feature, you will be asked to authorise Cambridge Core to connect with your <service> account. Find out more about sending content to Google Drive.

        Soya products and serum lipids: a meta-analysis of randomised controlled trials
        Available formats
Export citation


Soya proteins and isoflavones have been reported to exert beneficial effects on the serum lipid profile. More recently, this claim is being challenged. The objective of this study was to comprehensively examine the effects of soya consumption on the lipid profile using published trials. A detailed literature search was conducted via MEDLINE (from 2004 through February 2014), CENTRAL (The Cochrane Controlled Clinical Trials Register) and for randomised controlled trials assessing the effects of soya on the lipid profile. The primary effect measure was the difference in means of the final measurements between the intervention and control groups. In all, thirty-five studies (fifty comparisons) were included in our analyses. Treatment duration ranged from 4 weeks to 1 year. Intake of soya products resulted in a significant reduction in serum LDL-cholesterol concentration, –4·83 (95 % CI –7·34, –2·31) mg/dl, TAG, –4·92 (95 % CI –7·79, –2·04) mg/dl, and total cholesterol (TC) concentrations, –5·33 (95 % CI –8·35, –2·30) mg/dl. There was also a significant increase in serum HDL-cholesterol concentration, 1·40 (95 % CI 0·58, 2·23) mg/dl. The I 2 statistic ranged from 92 to 99 %, indicating significant heterogeneity. LDL reductions were more marked in hypercholesterolaemic patients, –7·47 (95 % CI –11·79, –3·16) mg/dl, than in healthy subjects, –2·96 (95 % CI –5·28, –0·65) mg/dl. LDL reduction was stronger when whole soya products (soya milk, soyabeans and nuts) were used as the test regimen, –11·06 (95 % CI –15·74, –6·37) mg/dl, as opposed to when ‘processed’ soya extracts, –3·17 (95 % CI –5·75, –0·58) mg/dl, were used. These data are consistent with the beneficial effects of soya proteins on serum LDL, HDL, TAG and TC concentrations. The effect was stronger in hypercholesterolaemic subjects. Whole soya foods appeared to be more beneficial than soya supplementation, whereas isoflavone supplementation had no effects on the lipid profile.

Dyslipidaemia is a primary risk factor for CVD, peripheral vascular disease and stroke. The WHO estimates that over 60 % of CHD and 40 % of ischaemic stroke in developed countries are due to total blood cholesterol levels in excess of the theoretical minimum of 3·8 mmol/l( 1 ). Reduction of serum LDL-cholesterol by about 5–6 %( 2 ) has the potential to reduce CHD risk by 7–12 %( 3 ), whereas a 3 % increase in HDL-cholesterol has the potential to lower the risk by 6–9 %( 4 , 5 ). Fasting serum TAG elevations also increase the risk for CHD( 6 ). The aggregate changes in these three lipoprotein risk factors can, therefore, potentially reduce CHD risk by 12–20 %( 7 ). As a result, cholesterol is by far the most studied risk factor for CHD risk( 8 ).

Although what constitutes an optimal dietary regimen remains a matter of controversy, current guidelines recommend diet as a first-line therapy for patients with elevated blood cholesterol concentrations( 9 ). Among the foods being suggested to lower blood lipids are soya products. A meta-analysis published in 1995 concluded that soya protein intake is effective in reducing total cholesterol (TC), LDL-cholesterol and TAG and in increasing HDL-cholesterol( 10 ). Following this, in October 1999, the US Food and Drug Administration approved labelling of foods containing soya protein as protective against CHD( 11 ); the American Heart Association released a similar advisory shortly thereafter( 12 ). The UK Joint Health Claims Initiative also permitted that ‘the inclusion of at least 25 g of soya protein/ d as part of a diet low in saturated fat can help reduce blood cholesterol’. Similar claims were also made in South Africa, the Philippines, Brazil and Indonesia( 13 ). All these presuppose that the hypolipidaemic effects of soya are due to its soya protein content. Soya products in addition, however, contain isoflavones, which have also been subjects of considerable scientific enquiry. Reports of trials investigating the effect of soya isoflavones on blood lipids have not been consistent( 14 ).

Over a 100 randomised controlled trials (RCT) and ten meta-analyses have already been published, which have examined the effects of soya protein and/or isoflavones on lipids. Most of these studies report varying effects of soya proteins and/or soya isoflavones on blood cholesterol. Perhaps the only consensus on the subject is that consumption of soya products may have a more marked hypolipidaemic effect in hypercholesterolaemic individuals than in normocholesterolaemic subjects. Although the ten previously completed meta-analyses have reported beneficial effects of soya protein and/or isoflavone on serum lipids, this favourable effect is being challenged, and a reason may be because some recently completed trials do not report an effect of soya protein on serum lipids. In 2012, the European Food Safety Authority concluded that ‘a cause and effect relationship has not been established between the consumption of isolated soya protein and a reduction in blood LDL-cholesterol concentrations’( 15 ). The last published meta-analysis involved studies completed between 1996 and 2008, and examined mainly the effects of soya proteins on LDL-cholesterol( 7 ).

In the past 6 years, nine additional clinical trials (involving twelve comparisons and 668 people) on the effects of soya products on lipids have been completed. Based on the availability of more recent high-quality trials, the objective of this review was to re-examine the conclusion that soya protein has proven beneficial effects on blood cholesterol. We also assessed the effects of isoflavones and the combined effect of soya protein and isoflavones on blood cholesterol levels. Specifically, we sought to answer the following question: Does habitual consumption of soya products have a beneficial effect on blood cholesterol level, and if it does what are the predictors of this effect?


Search strategy and study selection

On 3 March 2014, we searched MEDLINE (through February 2014), CENTRAL (The Cochrane Controlled Clinical Trials Register) and the website to identify RCT examining the effect of soya products on blood lipid profile. For the MEDLINE search, we used MeSH (major subject heading) terms (‘soybeans’ OR ‘soybean oil’ OR ‘soybean proteins’ OR ‘soy isoflavones’ OR ‘soy milk’) AND (‘Cholesterol, HDL’ OR ‘Cholesterol, LDL’ OR ‘Triglycerides’). The search was limited to the criteria ‘clinical trials’, ‘English’, ‘human’ and ‘last 10 years’. We limited to the last 10 years because the last published meta-analyses on this subject was completed in 2008. CENTRAL and the website were searched using ‘soy’ or ‘soybeans’ as free terms without further restrictions applied to the search. Inspection of the reference list of all identified articles was also conducted. The latter method was repeated until all potentially relevant articles from these sources were identified. Retrieved studies were included if they met the following criteria: (a) were investigating dietary soya products; (b) had a randomised controlled parallel-arm or cross-over design; (c) included subjects aged 18 years and above; (d) reported ‘end of intervention’ mean and standard deviation values of lipid measurements for the active (intervention) and control groups; and (e) provided the intake amount of soya products. We a priori determined that unpublished materials and conference abstracts will be excluded from the review. One-day trials were also excluded because we wanted to report the effects of habitual intake of soya products. Care was taken not to include data from multiple publications of the same population. Furthermore, we excluded studies in which soya intake was mixed with other dietary treatments.

In a sub-study, we searched MEDLINE (through March 2014), CENTRAL and the website to identify RCT examining the effect of soya isoflavone supplementation on blood lipid profile. We used the MeSH terms (‘soy isoflavone’ OR ‘daidzein’ OR ‘genistein’ OR ‘glycitein’) AND (‘Cholesterol, HDL’ OR ‘Cholesterol, LDL’ OR ‘Triglycerides’). The search was limited to the criteria ‘clinical trials’, ‘English’ and ‘human’.

Data extraction and quality assessment

Data were extracted by the lead author and subsequently reviewed by one of the co-authors for accuracy. Extracted data included the study characteristics (first author’s name; year of publication; number and age of participants; study design; daily amount of soya product consumed in the active arm; duration of the study; health characteristics of the study population; and location of the study). We also extracted information on the final concentrations of plasma/serum TC, LDL-cholesterol, HDL-cholesterol and TAG in both treatment arms. When the requisite data were not available in the published paper, authors were emailed requesting the appropriate information. All the authors agreed on the eligibility criteria of the included studies. Quality of the studies was evaluated using the validated Jadad score instrument( 16 ) with criteria that included the following: randomisation, adequacy of sequence generation, double blinding and description of drop-outs. The maximum score obtainable was 5.

Data synthesis and statistical analysis

The effect size used in this investigation was the ‘difference in means’ between the two treatment groups. We utilised plasma/serum cholesterol concentrations obtained at the end of each intervention. Ideally, serum cholesterol concentrations are approximately 3 % higher than the corresponding plasma concentrations( 17 ), but because we were interested in mean differences within each study we analysed plasma and serum concentrations without adjustment for this difference. Studies that reported results in mmol/l were converted to mg/dl using the standard conversion factors (which was a division of the mmol/l value by 0·02586 for TC, LDL and HDL and by 0·01129 for TAG). Estimates of treatment effect on cholesterol in sub-groups defined according to study design features were also carried out. Additional analyses were performed according to the Cochrane Handbook for Systematic Reviews of Interventions ( 18 ). Heterogeneity across studies was assessed by the Cochran’s Q test, and P<0·10 was considered statistically significant for heterogeneity. The magnitude of heterogeneity was evaluated by the I 2 statistic (percentage of the variability in effect estimates that is due to heterogeneity rather than sampling error). Whenever the test for heterogeneity was statistically significant, the estimate of the difference was calculated using the random effects model, according to DerSimonian & Laird( 19 ). For the computation of pooled effects, each study was assigned a weight consisting of the reciprocal of its variance. Furthermore, we performed a meta-regression to investigate whether there were any strong predictors of serum lipid changes. The included covariate variables were as follows: (1) initial lipid concentration; (2) duration of intake of soya products (weeks); (3) dose of soya protein consumed measured in grams per day; and (4) study quality (Jadad score). Each coefficient for the meta-regression analysis was estimated using restricted maximum likelihood and the corresponding variances were calculated using the ‘Knapp–Hartung’ variance estimator. Each meta-regression F test was conducted at the significance level of α=0·05, and all meta-regression analyses were performed using Stata/IC 13.1.

Data synthesis and other analyses were completed using Cochrane Collaboration Review Manager 5( 20 ) and Microsoft Office Excel 2010 package (Microsoft Corporation).


A total of sixty-two articles met our inclusion criteria and their complete texts were downloaded and fully reviewed. A flow chart depicting the selection process is depicted in Fig. 1. Two studies were excluded because there was an inappropriate control group( 21 , 22 ); three because they were 1-d trials( 23 25 ); five exclusions were due to the administration of a mixed test regimen( 26 30 ); and three studies were excluded due to lack of random assignment of the study groups( 31 33 ). Other studies were excluded for varying reasons (shown in Fig. 1). In all, thirty-five studies with fifty comparisons were included from our final analyses. Values used for the included studies represent the lipid measurements that correspond to the longest follow-up point. The observed I 2 statistic (percentage of the variability in effect estimates that is due to heterogeneity rather than sampling error) ranged from 92 to 99 %.

Fig. 1. Study selection process.

In the sub-study examining the effect of soya isoflavone supplementation on blood lipids, our search produced thirty-two results, including four of the studies already included in our primary study. After exclusion of the twenty-one ineligible studies, we had eleven studies( 34 44 ) with fifteen comparison groups.

Study and participant characteristics

The characteristics of the included studies are presented in Table 1. Duration of the studies ranged from 4 weeks to 1 year, with many lasting about 8 weeks. Most RCT were double-blind, described an adequate method of sequence generation and reported details of drop-outs and withdrawals. Overall, twenty-four of the thirty-five studies obtained a Jadad score of ≥3 (high quality).

Table 1 Characteristics of the included studies

A total of 2670 subjects (aged 28–83 years and 82 % women) were included in the primary analyses. The average intake of soya protein was 30 g/d (range: 14–50 g/d).

Effects of soya on lipid concentration

Intake of soya products resulted in a significant reduction in serum LDL-cholesterol concentration, –4·83 (95 % CI –7·34, –2·31) mg/dl (Fig. 2). The effect quality test (Cochran’s Q) did show significant heterogeneity, and thus the random effects model was used. We also observed statistically significant reduction in serum TAG, –4·92 (95 % CI –7·79, –2·04) mg/dl (Fig. 3), and TC concentrations, –5·33 (95 % CI –8·35, –2·30) mg/dl (Fig. 4). There was also a modest, but highly significant increase in serum HDL-cholesterol concentration, 1·40 (95 % CI 0·58, 2·23) mg/dl (Fig. 5).

Fig. 2. Meta-analysis of the effect of soya on LDL-cholesterol. The sizes of the data markers indicate the weight of each study in the analysis. IV, inverse variance. Random–random effects model. Values are in mg/dl.

Fig. 3. Meta-analysis of the effect of soya on serum TAG. The sizes of the data markers indicate the weight of each study in the analysis. IV, inverse variance. Random–random effects model. Values are in mg/dl.

Fig. 4. Meta-analysis of the effect of soya on total cholesterol. The sizes of the data markers indicate the weight of each study in the analysis. IV, inverse variance. Random–random effects model. Values are in mg/dl.

Fig. 5. Meta-analysis of the effect of soya on HDL-cholesterol. The sizes of the data markers indicate the weight of each study in the analysis. IV, inverse variance. Random–random effects model. Values are in mg/dl.

For the sub-study in which only isoflavone supplementation was administered as the test regimen, all the results (LDL, HDL, TAG and TC) were non-significant.

Sub-group analyses

LDL reductions also appeared to be more marked in hypercholesterolaemic patients, –7·47 (95 % CI –11·79, –3·16) mg/dl, than in healthy subjects, –2·96 (95 % CI –5·28, –0·65) mg/dl. When natural soya products (soya milk, whole soyabeans and soya nuts) were used as the test regimen, we observed a substantial reduction in serum LDL levels, –11·06 (95 % CI –15·74, –6·37) mg/dl, as opposed to when ‘processed’ soya (soya extracts, supplements), –3·17 (95 % CI –5·75, –0·58) mg/dl, was used. Intervention with soya protein alone seemed to exert a more positive effect on LDL and TC levels than intervention combining soya protein and isoflavone (Table 2). Studies adjudged to be of higher quality provided more precise estimates. Details of further analyses in sub-groups of studies defined by these study/participant characteristics are shown in Table 2. Sensitivity analyses conducted by excluding studies that used no control group( 45 ), used soya spread as the test diet( 46 ) or that did not have any washout period( 47 ) did not affect the results.

Table 2 Effects of soya proteins on serum lipids by study design, duration, location, health status and quality of study (Mean differences and 95 % confidence intervals)

* Significant results.

Isoflavone sub-study

The results of the sub-study that focused on the effect of isoflavone supplementation on LDL, HDL, TAG and TC were non-significant (Table 3). Sub-group analyses was also completed for each of these indices using design (cross-over v. parallel-arm), duration (≤12 v. >12 weeks) and health status (Table 3).

Table 3 Effects of soya isoflavone on serum lipids (Mean differences and 95 % confidence intervals)

* Significant results.

Predictors of the effect of soya on blood lipids

The meta-regression analysis suggested that initial serum LDL, TAG and TC concentrations are strong predictors of the effect of soya on serum LDL, TAG and TC concentrations, respectively (Table 4; Fig. 68). Initial HDL concentration did not predict the effect of soya on blood HDL levels. Dose of soya protein, the duration of its consumption and the study quality (Jadad score) also did not predict the effect of soya on blood lipid levels (Table 4).

Fig. 6. Meta-regression plot of the effect of initial LDL concentration on soya’s effect on LDL-cholesterol. Values are in mg/dl.

Fig. 7. Meta-regression plot of the effect of initial TAG concentration on soya’s effect on blood TAG level. Values are in mg/dl.

Fig. 8. Meta-regression plot of the effect of initial total concentration (TC) concentration on soya’s effect on blood TC level. Values are in mg/dl.

Table 4 Meta-regression – predictors of the effects of soya on serum lipids (Coefficients and their standard errors)

*Significant results.


Based on baseline lipid concentrations in the study population, our results show a significant 3 % reduction in serum LDL, 4 % reduction in serum TAG and a 2 % reduction in TC concentrations after an intervention with soya products for a period ranging from 4 to 52 weeks. We also observed a significant 3 % increase in serum HDL concentrations. This beneficial effect seemed stronger in individuals with a higher risk of CHD (hypercholesterolaemic, obese and diabetic subjects). It also appears that consumption of natural soya products is more effective in lowering serum cholesterol than intake of processed soya (e.g. soya protein extract preparations or supplements). LDL reduction was significant in the shorter-duration studies (4–8 weeks), whereas TAG reduction was only significant in the longer-lasting studies (10–52 weeks). Isoflavone supplementation only (i.e. without soya proteins) did not appear to significantly influence the serum lipid profile.

Various mechanisms have been suggested by which soya proteins may exert their lipid-lowering effects. Some have proposed the activation of LDL receptors by essential amino acids from soya protein( 48 ); others have discussed the possibility of a soya protein-based inhibition of endogenous cholesterol synthesis( 49 ). Water-soluble fibre and other components of soya may also reduce serum lipid levels( 12 ). Finally, it has been hypothesised that the substitution of soya for animal protein can result in lower saturated fat and cholesterol intakes, thereby indirectly resulting in a more favourable blood cholesterol levels( 50 ).

As casein has been reported to raise cholesterol levels, the relative increase in cholesterol concentration in the control group may be attributable to the casein content of the milk protein diet fed to the controls; however, when our analysis was restricted to the twenty-two comparisons where non-milk/non-casein-based diets were used as the control diet, the beneficial effect of the soya intervention on serum lipid was even more profound. Change in serum concentration was –6·37 (95 % CI –10·52, –2·22) mg/dl, –9·39 (95 % CI –12·88, –5·90) mg/dl and –9·69 (95 % CI –13·35, –6·04) mg/dl for LDL, TAG and TC, respectively.

Previous meta-analyses have reported on the hypolipidaemic effects of soya proteins( 7 , 51 , 52 ), soya isoflavones( 53 ) and soya proteins in combination with soya isoflavones( 14 ). Our results are consistent with analyses that report the cholesterol-lowering effects of soya protein. We, however, observed no effect of soya isoflavones on serum cholesterol levels and a reduced LDL-lowering effect in studies that administered soya proteins in combination with isoflavones (compared with studies that administered soya proteins alone). Soya protein plus isoflavone test diets, however, had a more positive effect on serum TAG concentrations.

Mean daily intake of soya protein is about 30 g in Japan, 20 g in Korea, 8 g in China and <1 g in the USA( 54 , 55 ). The mean daily intake by participants in our analyses was 29·8 g (range: 14–50 g). This potentially means that the cholesterol-reducing effects of soya protein may not be realised by most soya consumers. It is important to note, however, that a dose–response effect of soya protein on cholesterol reduction was not observed in our analyses and is yet to be established.

As soyabeans are grown around the world under many different climatic conditions, there is a wide range of soyabean varieties. Unlike in Europe, GM soyabeans have become the predominant type grown in the Americas. When our results were stratified by study location, the studies completed in Europe, Asia and Australia showed highly significant positive effects. Conversely, the studies completed in North/South America (where >85 % of soyabean produced is GM) showed non-significant results, except for a mild elevation of serum HDL.

In an earlier study, we observed varying effects of flavanols on serum lipids, depending on the matrix with which the flavanols were administered( 56 ). Consequently, in this analysis, we compared studies that used natural soya products (roasted soyabeans, soya nuts, etc.) with those that used soya extracts or tablets as the test regimen (Table 2). The LDL-cholesterol-lowering effect appeared to be three times stronger when natural soya products were used. A potential explanation may be the differing levels of bioavailability of the active ingredients.

Soya isoflavones have a biological similarity to mammalian oestrogens, which have been shown to exert cholesterol-lowering effects in humans( 57 ). Our soya isoflavone sub-study revealed no effects of isoflavone supplementation on serum lipids. In their meta-analysis, Weggemans & Trautwein( 58 ) also reported a lack of efficacy of soya isoflavones on changes in LDL or HDL-cholesterol. In contrast, another study has reported positive effects of isoflavone supplementation on cholesterol and endothelial function( 39 ). The varying conclusions on the effect of isoflavones may be in part due to the process by which the isoflavones were extracted( 59 ). Mean isoflavone consumption is 11–47 mg/d in Asian countries and 1–2 mg/d in Western countries( 60 , 61 ), but the mean daily isoflavone consumption in the included studies was 76 mg.

Although we observed a high level of heterogeneity, we are still confident in the conclusions of these analyses, as our observations are consistent with many of the other previously published reports.

In conclusion, our findings showed that an intervention with soya proteins increases serum HDL concentration and lowers serum TAG, LDL and TC concentrations. No effect of isoflavones was observed on serum lipids, and natural soya products appear to have a stronger hypolipidaemic effect on serum cholesterol than soya supplements. Finally, initial LDL, TAG and TC concentrations seem to be strong predictors of the effect of soya on blood lipid levels.


During the last 3 years, O. A. T. reports that he has received an investigator-initiated grant from the National Institutes of Health; L. D. reports that he has received investigator-initiated grants from the National Institutes of Health, GlaxoSmithKline, Amarin Pharma Inc., California Walnut Commission and Merck, and has received travel re-imbursement from the International Nut & Dried Fruit Council Inc.; J. M. G. reports that he has received investigator-initiated grants from the National Institutes of Health, the Veterans Administration, Amgen and pills and packaging from Pfizer for a research study, and has served as a consultant to Bayer.

L. D. and O. A. T. conceived the study and are responsible for data integrity and accuracy of data analyses. O. A. T. and T. A. O. drafted the manuscript, and along with L. D. have the primary responsibility for final content. O. A. T., A. Y. and T. A. O. completed the statistical analyses; J. M. G. reviewed the manuscript for its scientific content. All the authors have read and approved this final draft.

The authors have no conflict of interest to declare.


1. World Health Organization (2002) The World Health Report 2002; Reducing Risks, Promoting Healthy Life. Geneva: WHO.
2. Sirtori, CR, Anderson, JW & Arnoldi, A (2007) Nutritional and nutraceutical considerations for dyslipidemia. Future Med 2, 313339.
3. Katan, MB, Grundy, SM, Jones, P, et al. (2003) Efficacy and safety of plant stanols and sterols in the management of blood cholesterol levels. Mayo Clin Proc 78, 965978.
4. Anderson, JW & Konz, EC (2001) Obesity and disease management: effects of weight loss on comorbid conditions. Obes Res 9, 326S334S.
5. Ashen, MD & Blumenthal, RS (2005) Low HDL cholesterol levels. N Engl J Med 353, 12521260.
6. Ginsberg, HN (1997) Is hypertriglyceridemia a risk factor for atherosclerotic cardiovascular disease? A simple question with a complicated answer. Ann Intern Med 126, 912914.
7. Anderson, JW & Bush, HM (2011) Soy protein effects on serum lipoproteins: a quality assessment and meta-analysis of randomized, controlled studies. J Am Coll Nutr 30, 7991.
8. Sacks, FM, Lichtenstein, A, Van Horn, L, et al. (2006) Soy protein, isoflavones, and cardiovascular health: an American Heart Association Science Advisory for professionals from the Nutrition Committee. Circulation 113, 10341044.
9. Ferdowsian, HR & Barnard, ND (2009) Effects of plant-based diets on plasma lipids. Am J Cardiol 104, 947956.
10. Anderson, JW, Johnstone, BM & Cook-Newell, ME (1995) Meta-analysis of the effects of soy protein intake on serum lipids. N Engl J Med 333, 276282.
11. Food & Administration D (1999) Food labeling health claims; soy protein and coronary heart disease. Fed Regist 64, 5769957733.
12. Erdman, JW (2000) Soy protein and cardiovascular disease: a statement for healthcare professionals from the Nutrition Committee of the AHA. Circulation 102, 25552559.
13. Xiao, CW (2008) Health effects of soy protein and isoflavones in humans. J Nutr 138, 1244S1249S.
14. Zhan, S & Ho, SC (2005) Meta-analysis of the effects of soy protein containing isoflavones on the lipid profile. Am J Clin Nutr 81, 397408.
15. EFSA Panel on Dietetic Products Nutrition and Allergies (2012) Scientific opinion on the substantiation of a health claim related to isolated soy protein and reduction of blood LDL-cholesterol concentrations pursuant to Article 14 of Regulation (EC) No 1924/2006. EFSA Journal 10, 2555.
16. Jadad, AR, Moore, RA, Carroll, D, et al. (1996) Assessing the quality of reports of randomized clinical trials: is blinding necessary? Control Clin Trials 17, 112.
17. Lipid Research Clinics Laboratory Methods Committee (1977) Cholesterol and triglyceride concentrations in serum/plasma pairs. Clin Chem 23, 6063.
18. Higgins, J & Green, S (2010) Cochrane Handbook for Systematic Reviews of Interventions Version 5.0. 2. The Cochrane Collaboration, 2009. Chichester: John Wiley.
19. DerSimonian, R & Laird, N (1986) Meta-analysis in clinical trials. Control Clin Trials 7, 177188.
20. Collaboration RTC (2008) Review Manager (RevMan). 5.0. Copenhagen, The Nordic Cochrane Centre, The Cochrane Collaboration.
21. Engelman, HM, Alekel, DL, Hanson, LN, et al. (2005) Blood lipid and oxidative stress responses to soy protein with isoflavones and phytic acid in postmenopausal women. Am J Clin Nutr 81, 590596.
22. Ahn-Jarvis, J, Clinton, SK, Riedl, KM, et al. (2012) Impact of food matrix on isoflavone metabolism and cardiovascular biomarkers in adults with hypercholesterolemia. Food Funct 3, 10511058.
23. Campbell, CG, Brown, BD, Dufner, D, et al. (2006) Effects of soy or milk protein during a high-fat feeding challenge on oxidative stress, inflammation, and lipids in healthy men. Lipids 41, 257265.
24. Hall, WL, Formanuik, NL, Harnpanich, D, et al. (2008) A meal enriched with soy isoflavones increases nitric oxide-mediated vasodilation in healthy postmenopausal women. J Nutr 138, 12881292.
25. Hanwell, HE, Kay, CD, Lampe, JW, et al. (2009) Acute fish oil and soy isoflavone supplementation increase postprandial serum (n-3) polyunsaturated fatty acids and isoflavones but do not affect triacylglycerols or biomarkers of oxidative stress in overweight and obese hypertriglyceridemic men. J Nutr 139, 11281134.
26. Matthan, NR, Welty, FK, Barrett, PHR, et al. (2004) Dietary hydrogenated fat increases high-density lipoprotein apoA-I catabolism and decreases low-density lipoprotein apoB-100 catabolism in hypercholesterolemic women. Arterioscler Thromb Vasc Biol 24, 10921097.
27. Jenkins, DJ, Kendall, CW, Marchie, A, et al. (2005) Direct comparison of a dietary portfolio of cholesterol-lowering foods with a statin in hypercholesterolemic participants. Am J Clin Nutr 81, 380387.
28. Verhoeven, MO, Teerlink, T, Kenemans, P, et al. (2007) Effects of a supplement containing isoflavones and Actaea racemosa L. on asymmetric dimethylarginine, lipids, and C-reactive protein in menopausal women. Fertil Steril 87, 849857.
29. Weidner, C, Krempf, M, Bard, JM, et al. (2008) Cholesterol lowering effect of a soy drink enriched with plant sterols in a French population with moderate hypercholesterolemia. Lipids Health Dis 7, 3542.
30. Rideout, TC, Chan, Y-M, Harding, SV, et al. (2009) Low and moderate-fat plant sterol fortified soymilk in modulation of plasma lipids and cholesterol kinetics in subjects with normal to high cholesterol concentrations: report on two randomized crossover studies. Lipids Health Dis 8, 45.
31. Jenkins, DJ, Kendall, CW, Nguyen, TH, et al. (2008) Effect of plant sterols in combination with other cholesterol-lowering foods. Metabolism 57, 130139.
32. Torres, N, Guevara-Cruz, M, Granados, J, et al. (2009) Reduction of serum lipids by soy protein and soluble fiber is not associated with the ABCG5/G8, apolipoprotein E, and apolipoprotein A1 polymorphisms in a group of hyperlipidemic Mexican subjects. Nutr Res 29, 728735.
33. Lobato, LP, Iakmiu Camargo Pereira, AE, Lazaretti, MM, et al. (2012) Snack bars with high soy protein and isoflavone content for use in diets to control dyslipidaemia. Int J Food Sci Nutr 63, 4958.
34. Hodgson, JM, Puddey, IB, Beilin, LJ, et al. (1998) Supplementation with isoflavonoid phytoestrogens does not alter serum lipid concentrations: a randomized controlled trial in humans. J Nutr 128, 728732.
35. Merz-Demlow, BE, Duncan, AM, Wangen, KE, et al. (2000) Soy isoflavones improve plasma lipids in normocholesterolemic, premenopausal women. Am J Clin Nutr 71, 14621469.
36. Wangen, KE, Duncan, AM, Xu, X, et al. (2001) Soy isoflavones improve plasma lipids in normocholesterolemic and mildly hypercholesterolemic postmenopausal women. Am J Clin Nutr 73, 225231.
37. Uesugi, T, Fukui, Y & Yamori, Y (2002) Beneficial effects of soybean isoflavone supplementation on bone metabolism and serum lipids in postmenopausal Japanese women: a four-week study. J Am Coll Nutr 21, 97102.
38. Cheng, SY, Shaw, NS, Tsai, KS, et al. (2004) The hypoglycemic effects of soy isoflavones on postmenopausal women. J Womens Health 13, 10801086.
39. Colacurci, N, Chiantera, A, Fornaro, F, et al. (2005) Effects of soy isoflavones on endothelial function in healthy postmenopausal women. Menopause 12, 299307.
40. Garrido, A, la Maza, D, Pia, M, et al. (2006) Soy isoflavones affect platelet thromboxane A2 receptor density but not plasma lipids in menopausal women. Maturitas 54, 270276.
41. Hall, WL, Vafeiadou, K, Hallund, J, et al. (2006) Soy-isoflavone-enriched foods and markers of lipid and glucose metabolism in postmenopausal women: interactions with genotype and equol production. Am J Clin Nutr 83, 592600.
42. Ho, SC, Chen, YM, Ho, SS, et al. (2007) Soy isoflavone supplementation and fasting serum glucose and lipid profile among postmenopausal Chinese women: a double-blind, randomized, placebo-controlled trial. Menopause 14, 905912.
43. Rios, DRA, Rodrigues, ET, Cardoso, APZ, et al. (2008) Lack of effects of isoflavones on the lipid profile of Brazilian postmenopausal women. Nutrition 24, 11531158.
44. Qin, Y, Shu, F, Zeng, Y, et al. (2014) Daidzein supplementation decreases serum triglyceride and uric acid concentrations in hypercholesterolemic adults with the effect on triglycerides being greater in those with the GA compared with the GG genotype of ESR-β RsaI. J Nutr 144, 4954.
45. Bakhtiary, A, Yassin, Z, Hanachi, P, et al. (2012) Effects of soy on metabolic biomarkers of cardiovascular disease in elderly women with metabolic syndrome. Arch Iran Med 15, 462468.
46. Clifton, P, Mano, M, Duchateau, G, et al. (2008) Dose-response effects of different plant sterol sources in fat spreads on serum lipids and C-reactive protein and on the kinetic behavior of serum plant sterols. Eur J Clin Nutr 62, 968977.
47. Meyer, BJ, Larkin, TA, Owen, AJ, et al. (2004) Limited lipid-lowering effects of regular consumption of whole soybean foods. Ann Nutr Metab 48, 6778.
48. Gianazza, E, Lovati, M, Manzoni, C, et al. (1998) Reduction of serum cholesterol by soy proteins: clinical experience and potential molecular mechanisms. Nutr Metab Cardiovasc Dis 8, 334340.
49. Tham, DM, Gardner, CD & Haskell, WL (1998) Potential health benefits of dietary phytoestrogens: a review of the clinical, epidemiological, and mechanistic evidence 1. J Clin Endocrinol Metab 83, 22232235.
50. Lichtenstein, AH (1998) Soy protein, isoflavones and cardiovascular disease risk. J Nutr 128, 15891592.
51. Reynolds, K, Chin, A, Lees, KA, et al. (2006) A meta-analysis of the effect of soy protein supplementation on serum lipids. Am J Cardiol 98, 633640.
52. Harland, JI & Haffner, TA (2008) Systematic review, meta-analysis and regression of randomised controlled trials reporting an association between an intake of circa 25g soya protein per day and blood cholesterol. Atherosclerosis 200, 1327.
53. Taku, K, Umegaki, K, Sato, Y, et al. (2007) Soy isoflavones lower serum total and LDL cholesterol in humans: a meta-analysis of 11 randomized controlled trials. Am J Clin Nutr 85, 11481156.
54. Ho, SC, Woo, JL, Leung, SS, et al. (2000) Intake of soy products is associated with better plasma lipid profiles in the Hong Kong Chinese population. J Nutr 130, 25902593.
55. Nagata, C (2000) Ecological study of the association between soy product intake and mortality from cancer and heart disease in Japan. Int J Epidemiol 29, 832836.
56. Tokede, O, Gaziano, J & Djoussé, L (2011) Effects of cocoa products/dark chocolate on serum lipids: a meta-analysis. Eur J Clin Nutr 65, 879886.
57. Rossouw, JE (1999) Hormone replacement therapy and cardiovascular disease. Curr Opin Lipidol 10, 429434.
58. Weggemans, R & Trautwein, E (2003) Relation between soy-associated isoflavones and LDL and HDL cholesterol concentrations in humans: a meta-analysis. Eur J Clin Nutr 57, 940946.
59. Clarkson, TB & Anthony, MS (1998) Phytoestrogens and coronary heart disease. Baillieres Clin Endocrinol Metab 12, 589604.
60. Arai, Y, Watanabe, S, Kimira, M, et al. (2000) Dietary intakes of flavonols, flavones and isoflavones by Japanese women and the inverse correlation between quercetin intake and plasma LDL cholesterol concentration. J Nutr 130, 22432250.
61. de Kleijn, MJ, van der Schouw, YT, Wilson, PW, et al. (2001) Intake of dietary phytoestrogens is low in postmenopausal women in the United States: the Framingham study (1–4). J Nutr 131, 18261832.