CHD is the leading cause of morbidity and mortality in the USA( Reference Dalen and Devries 1 ) and worldwide( 2 ). Nutrition is an important determinant for the risk of developing CHD; poor dietary habits are estimated to account for 20 % of CHD cases in the US adult population( Reference Dalen and Devries 1 ). Soya foods are a potential nutritional source for modifying biomarkers of CHD( Reference Setchell, Brown and Lydeking-Olsen 3 , Reference Matori, Umar and Nadadur 4 ). One of the main components of soya that may exert protective cardioprotective effects are isoflavones, bioactive phyto-oestrogens found in soyabeans( Reference Setchell, Brown and Lydeking-Olsen 3 ). The predominant soya isoflavones are genistein, daidzein and glycitein. Isoflavones may reduce the risk of CHD by: (1) their action via oestrogen receptor β, due to their structural similarity to oestradiol, leading to decreased vasodilation and inflammation( Reference Matori, Umar and Nadadur 4 – Reference Mahmoud, Yang and Bosland 7 ); (2) their antioxidant activity, which may prevent the oxidative damage to LDL-cholesterol (LDL-C) that contributes to atherogenesis( Reference Lukito 8 ); and (3) modulating the vascular system, reducing atherosclerotic lesions and improving vascular reactivity and vascular stiffness( Reference Cano, Garcia-Perez and Tarin 9 , Reference Erdman 10 ).
Although there are clear cardiovascular benefits of isoflavones in vitro and in animal studies( Reference Cano, Garcia-Perez and Tarin 9 , Reference Mann, Bonacasa and Ishii 11 ), the evidence in humans is conflicting( Reference Zhan and Ho 12 – Reference Hodis, Mack and Kono 14 ). A growing hypothesis is that the ability of humans to metabolise daidzein to equol, referred to as ‘equol producers’, may contribute to the protective effects of soya( Reference Setchell and Clerici 15 , Reference Rowland, Wiseman and Sanders 16 ). Equol has a greater affinity for oestrogen receptors than its precursor daidzein( Reference Muthyala, Ju and Sheng 17 ), a longer half-life and bioavailability in plasma than daidzein and genistein( Reference Setchell, Brown and Lydeking-Olsen 3 , Reference Kelly, Joannou and Reeder 18 ), and more potent antioxidant activity than any other isoflavone( Reference Setchell, Brown and Lydeking-Olsen 3 ). Therefore, the potential beneficial effects of soya isoflavones for CHD and its risk factors may be greater among equol producers. While all tested animals, including rodents and monkeys, can produce equol, not all humans have the gut microflora required to convert daidzein to equol, a bioactive metabolite( Reference Setchell and Clerici 15 , Reference Atkinson, Frankenfeld and Lampe 19 ).
Equol is a promising candidate for hindering the initiation and progression of atherosclerosis due to its ability to induce vasorelaxation and its anti-inflammatory and antioxidant activity( Reference Setchell and Clerici 20 ). Specifically, it induces vasorelaxation through enhancing the production of endothelium nitric oxide synthase-derived NO( Reference Ohkura, Obayashi and Yamada 21 ). It can also inhibit NO derived by inducible nitric oxide synthase, expressed by immune cells during host defence, which is linked to atherosclerosis development( Reference Kang, Yoon and Han 22 ). Furthermore, equol prevents lipid and lipoprotein peroxidation, a crucial process in the pathogenesis of atherosclerosis( Reference Arora, Nair and Strasburg 23 , Reference Hodgson, Croft and Puddey 24 ).
The purpose of the present review is to examine if there is a difference in the cardioprotective effect of soya isoflavones in humans based on the hosts’ ability to produce equol. No previous reviews have thoroughly examined the impact of equol-producing status on risk factors for CHD. We conducted a comprehensive search of the scientific literature to identify randomised controlled trials (RCT) that evaluated the effects of soya isoflavones on risk factors for CHD and selected studies that included analyses based on equol producer status.
Methods
Literature search
The systematic review was conducted following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines( Reference Liberati, Altman and Tetzlaff 25 ). We initially searched PubMed (1950 to April 2015), EMBASE through Embase.com (1966 to April 2015), Ovid Medline (1946 to April 2015) and the Cochrane Library (Cochrane Central Register of Controlled Trials, 1999 to April 2015) for papers in any language using one or more textual or medical subject heading (MESH) terms for isoflavones (isoflavones, isoflavonoids, genistein, daidzein, equol), risk factors for CHD (cardiovascular disease, coronary heart disease, myocardial infarction, lipids, low-density lipoprotein-cholesterol, triglyceride, lipoproteins, hypercholesterolemia, lipid metabolism, blood pressure, glucose, vital signs, arterial stiffness, vascular stiffness, intima-media thickness, inflammation, endothelial function, endothelium, adipocytes) and RCT (randomised control study, clinical trial, placebo, intervention studies, pilot projects, sampling studies, twin studies, prospective studies, double blind study, single blind study, epidemiologic research design). We reviewed the reference lists of the collected articles to identify additional potentially relevant papers not identified by the original keyword search.
Study selection
Studies were selected for the systematic review if they met the following criteria: (1) RCT; (2) full-text was published in English; (3) analysed adult subjects who ingested soya with isoflavones or isolated isoflavones as an intervention; (4) analysed traditional risk factors for CHD (including lipids, inflammatory, blood pressure, glycaemic and body composition variables) as outcome measurements; (5) determined the equol producer status of the participants; and (6) stratified the outcome measurements by equol producer status. The exclusion criteria included reviews or commentaries.
Data synthesis and quality assessment
Searching, data extraction and the quality assessment were completed by two authors independently according to the inclusion criteria. Discrepancies were resolved by consensus. For each RCT, extracted data included sample size, baseline characteristics of the participants (sex, mean age, health status, demographics, equol producer status), study design, treatment regimen (dose, duration, isoflavone content, and type of soya intervention), and the assessment of the risk factor(s) for CHD.
The quality of the RCT methodology was graded using a fourteen-point evaluation tool for controlled clinical trials developed by the National Heart, Lung, and Blood Institute( 26 ). Questions were answered with a ‘yes’, ‘no’, ‘not reported’, ‘cannot determine’ or ‘not applicable’. The evaluation was based on the primary outcome measurements of the RCT. The RCT were given an overall rating of ‘good’, ‘fair’ or ‘poor’ at the discretion of the reviewers based on the guidelines provided by this tool.
Results
Search results
A total of 1671 papers were collected and, of these, 829 were excluded because they were not RCT, did not measure the traditional risk factors for CHD, or were not published in English (Fig. 1). Of the remaining 247 papers screened, forty-two met the selection criteria for this review. An outline of our search strategy using PubMed is provided in Supplementary Table S1.
Fig. 1. Study flow diagram of screened, excluded and analysed publications.
Study characteristics
Study characteristics are summarised in Tables 1 and 2, and Supplementary Table S2. Thirty studies included only female participants( Reference Hodis, Mack and Kono 14 , Reference Acharjee, Zhou and Elajami 27 – Reference Welty, Lee and Lew 55 ), eleven studies included both males and females( Reference Clerici, Setchell and Battezzati 56 – Reference Wong, Kendall and Marchie 66 ), and one study had only male participants( Reference McVeigh, Dillingham and Lampe 67 ). Of the forty-one studies involving women, thirty-four included postmenopausal women only( Reference Hodis, Mack and Kono 14 , Reference Acharjee, Zhou and Elajami 27 , Reference Badeau, Jauhiainen and Metso 28 , Reference Curtis, Potter and Kroon 30 – Reference Welty, Lee and Lew 55 , Reference Nestel, Cehun and Chronopoulos 60 , Reference Reverri, LaSalle and Franke 63 , Reference West, Hilpert and Juturu 65 , Reference Wong, Kendall and Marchie 66 ). The age of the participants ranged from 27 to 73 years. Participants were hypercholesterolaemic in seven studies( Reference Clerici, Setchell and Battezzati 56 , Reference Gardner, Messina and Kiazand 57 , Reference Meyer, Larkin and Owen 59 , Reference Qin, Shu and Zeng 62 , Reference Thorp, Howe and Mori 64 – Reference Wong, Kendall and Marchie 66 ), hyperlipidaemic in two studies( Reference Ma, Chiriboga and Olendzki 58 , Reference Wong, Kendall and Marchie 66 ), prehypertensive or hypertensive in five studies( Reference Liu, Ho and Chen 39 – Reference Liu, Ho and Chen 41 , Reference Welty, Lee and Lew 55 , Reference Meyer, Larkin and Owen 59 ), had type 2 diabetes in two studies( Reference Curtis, Potter and Kroon 30 , Reference Pipe, Gobert and Capes 61 ), had the metabolic syndrome in two studies( Reference Acharjee, Zhou and Elajami 27 , Reference Reverri, LaSalle and Franke 63 ), and considered healthy in twenty-three studies( Reference Hodis, Mack and Kono 14 , Reference Campbell, Woodside and Honour 29 , Reference Gallagher, Satpathy and Rafferty 31 , Reference Hall, Vafeiadou and Hallund 34 – Reference Kreijkamp-Kaspers, Kok and Grobbee 38 , Reference Mangano, Hutchins-Wiese and Kenny 42 – Reference van der Velpen, Geelen and Schouten 54 , Reference Nestel, Cehun and Chronopoulos 60 , Reference McVeigh, Dillingham and Lampe 67 ). Diet interventions in nineteen studies used soya protein isolate with isoflavone flour, or powder, or tablets( Reference Gallagher, Satpathy and Rafferty 31 – Reference Greany, Nettleton and Wangen 33 , Reference Kreijkamp-Kaspers, Kok and Bots 37 – Reference Mangano, Hutchins-Wiese and Kenny 42 , Reference Steinberg, Guthrie and Villablanca 48 – Reference Törmälä, Appt and Clarkson 51 , Reference Gardner, Messina and Kiazand 57 , Reference Ma, Chiriboga and Olendzki 58 , Reference Pipe, Gobert and Capes 61 , Reference Qin, Shu and Zeng 62 , Reference West, Hilpert and Juturu 65 , Reference McVeigh, Dillingham and Lampe 67 ), fifteen used soya- and isoflavone-enriched milk or foods( Reference Hodis, Mack and Kono 14 , Reference Acharjee, Zhou and Elajami 27 , Reference Curtis, Potter and Kroon 30 – Reference Hallund, Bugel and Tholstrup 36 , Reference Reimann, Dierkes and Carlsohn 46 , Reference Sen, Morimoto and Heak 47 , Reference Welty, Lee and Lew 55 – Reference Gardner, Messina and Kiazand 57 , Reference Meyer, Larkin and Owen 59 , Reference Reverri, LaSalle and Franke 63 – Reference Wong, Kendall and Marchie 66 ), and nine used isolated isoflavone tablets or capsules( Reference Badeau, Jauhiainen and Metso 28 , Reference Campbell, Woodside and Honour 29 , Reference Mangano, Hutchins-Wiese and Kenny 42 – Reference Pusparini and Hidayat 45 , Reference Törmälä, Nikander and Tiitinen 52 – Reference van der Velpen, Geelen and Schouten 54 ), with Gardner et al. ( Reference Gardner, Messina and Kiazand 57 ) using interventions that covered two categories. Isoflavone doses ranged from approximately 40 to 120 mg/d, with one dose particularly high at 900 mg/d( Reference Pop, Fischer and Coan 44 ). Twenty-three studies examined cholesterol markers( Reference Acharjee, Zhou and Elajami 27 , Reference Badeau, Jauhiainen and Metso 28 , Reference Gallagher, Satpathy and Rafferty 31 , Reference Greany, Nettleton and Wangen 32 , Reference Hall, Vafeiadou and Hallund 35 , Reference Kreijkamp-Kaspers, Kok and Grobbee 38 , Reference Liu, Ho and Chen 39 , Reference Mangano, Hutchins-Wiese and Kenny 42 , Reference Nikander, Tiitinen and Laitinen 43 , Reference Steinberg, Guthrie and Villablanca 48 , Reference Törmälä, Nikander and Tiitinen 52 , Reference Clerici, Setchell and Battezzati 56 – Reference McVeigh, Dillingham and Lampe 67 ), twenty-one examined other lipid variables( Reference Acharjee, Zhou and Elajami 27 , Reference Badeau, Jauhiainen and Metso 28 , Reference Gallagher, Satpathy and Rafferty 31 , Reference Greany, Nettleton and Wangen 32 , Reference Hall, Vafeiadou and Hallund 35 , Reference Kreijkamp-Kaspers, Kok and Grobbee 38 , Reference Liu, Ho and Chen 39 , Reference Mangano, Hutchins-Wiese and Kenny 42 , Reference Nikander, Tiitinen and Laitinen 43 , Reference Sen, Morimoto and Heak 47 , Reference Steinberg, Guthrie and Villablanca 48 , Reference Törmälä, Nikander and Tiitinen 52 , Reference Clerici, Setchell and Battezzati 56 , Reference Ma, Chiriboga and Olendzki 58 , Reference Meyer, Larkin and Owen 59 , Reference Pipe, Gobert and Capes 61 , Reference Qin, Shu and Zeng 62 , Reference Thorp, Howe and Mori 64 – Reference McVeigh, Dillingham and Lampe 67 ), eighteen examined blood pressure and vascular variables( Reference Hodis, Mack and Kono 14 , Reference Acharjee, Zhou and Elajami 27 , Reference Curtis, Potter and Kroon 30 , Reference Hall, Vafeiadou and Hallund 34 – Reference Kreijkamp-Kaspers, Kok and Bots 37 , Reference Liu, Ho and Chen 39 , Reference Liu, Ho and Chen 40 , Reference Nikander, Tiitinen and Laitinen 43 , Reference Steinberg, Guthrie and Villablanca 48 , Reference Törmälä, Appt and Clarkson 49 , Reference Törmälä, Appt and Clarkson 51 , Reference Welty, Lee and Lew 55 , Reference Clerici, Setchell and Battezzati 56 , Reference Meyer, Larkin and Owen 59 , Reference Reverri, LaSalle and Franke 63 , Reference Wong, Kendall and Marchie 66 ), seventeen examined inflammatory markers( Reference Acharjee, Zhou and Elajami 27 , Reference Greany, Nettleton and Wangen 33 , Reference Hall, Vafeiadou and Hallund 34 , Reference Mangano, Hutchins-Wiese and Kenny 42 , Reference Pop, Fischer and Coan 44 – Reference Reimann, Dierkes and Carlsohn 46 , Reference Steinberg, Guthrie and Villablanca 48 , Reference Törmälä, Appt and Clarkson 50 , Reference van der Velpen, Geelen and Hollman 53 , Reference van der Velpen, Geelen and Schouten 54 , Reference Clerici, Setchell and Battezzati 56 , Reference Qin, Shu and Zeng 62 , Reference Reverri, LaSalle and Franke 63 , Reference West, Hilpert and Juturu 65 – Reference McVeigh, Dillingham and Lampe 67 ), ten examined glucose and insulin variables( Reference Acharjee, Zhou and Elajami 27 , Reference Campbell, Woodside and Honour 29 , Reference Hall, Vafeiadou and Hallund 35 , Reference Liu, Ho and Chen 39 , Reference Nikander, Tiitinen and Laitinen 43 , Reference Törmälä, Appt and Clarkson 50 , Reference Gardner, Messina and Kiazand 57 , Reference Qin, Shu and Zeng 62 , Reference Reverri, LaSalle and Franke 63 , Reference West, Hilpert and Juturu 65 ) and five examined body composition variables( Reference Acharjee, Zhou and Elajami 27 , Reference Liu, Ho and Chen 41 , Reference Nikander, Tiitinen and Laitinen 43 , Reference West, Hilpert and Juturu 65 , Reference Wong, Kendall and Marchie 66 ). There were numerous methods and standards used to distinguish equol producers from non-equol producers, including sampling from urine or serum, different threshold levels for differentiation, and various analytical techniques.
Table 1. Demographic and clinical characteristics of the participants in the randomised controlled trials (RCT) employing soya interventions and examining the effect of equol producer (EP) status on risk factors for CHD
M, male; F, female; NEP, non-equol producer; MetS, metabolic syndrome; CO, crossover; DB, double-blind; P, parallel; B, blinded; SB, single-blinded; HRT, hormone replacement therapy.
* The quality of the RCT were evaluated based on the main outcomes reported. RCT were given a score of ‘good’, ‘fair’ or ‘poor’ after appraising the degree to which flaws in the study designs could affect the validity of the results.
† Studies that are or potentially using shared study participants.
‡ Studies that are or potentially using shared study participants.
§ Studies that are or potentially using shared study participants.
‖ Studies that are or potentially using shared study participants.
¶ Studies that are or potentially using shared study participants.
** Studies that are or potentially using shared study participants.
†† Studies that are or potentially using shared study participants.
Table 2. Characteristics of the soya isoflavone interventions used in the randomised controlled trials examining the effect of equol producer (EP) status on the risk factors for CHD
TLC, therapeutic lifestyle changes; AG, aglycone; SPI, soya protein isolate; WB, whole bean soya; MPI, milk protein isolate; N/A, not applicable.
* Studies that are or potentially using shared study participants.
† Studies that are or potentially using shared study participants.
‡ Studies that are or potentially using shared study participants.
§ Studies that are or potentially using shared study participants.
‖ Studies that are or potentially using shared study participants.
¶ Studies that are or potentially using shared study participants.
** Studies that are or potentially using shared study participants.
Synthesis of results
We categorised both the effects of soya isoflavones and equol producer status on the examined CHD risk factors as beneficial, negligible, or adverse (Tables 3–8). We analysed each risk factor independently; therefore the RCT were potentially categorised more than once. Twenty-two studies found statistically significant improvements in the risk factors for CHD after the soya isoflavone intervention compared with placebo. Of these, equol producer status further improved risk factors for CHD in six studies (including LDL-C, TAG, systolic blood pressure, diastolic blood pressure, flow-mediated dilation, soluble intercellular adhesion molecule-1, platelet-selectin and C-reactive protein). Equol producer status was comparable to the soya intervention in sixteen studies (including total cholesterol, LDL-C, HDL-cholesterol (HDL-C), TAG, apoB, systolic blood pressure, diastolic blood pressure, nitrate and nitrite, systemic arterial compliance, peak flow velocity, aortic augmentation index and IL-6).
Table 3. Randomised clinical trial results reporting the effect of soya isoflavone interventions and equol producer (EP) status on cholesterol and other lipid parameters*
MetS, metabolic syndrome; NEP, non-equol producer; LDL-C, LDL-cholesterol; sdLDL-C, small dense LDL-C; Lp, lipoprotein; TC, total cholesterol; HDL-C, HDL-cholesterol; WB, whole bean soya; SPI, soya protein isolate; ABCA1, adenosine triphosphate-binding cassette A1; CD, conjugated diene formation; OxLDL-C, oxidised LDL-C.
* Results are first stratified by the impact of EP status and then the impact of the soya isoflavone interventions on each of the lipid risk factors.
† +, Beneficial effect of soya isoflavones on risk factors of CHD; 0, negligible effect of soya isoflavones on risk factors of CHD; –, adverse effect of soya isoflavones on risk factors of CHD.
‡ +, Beneficial effect of EP status on risk factors of CHD after soya intervention; 0, negligible effect of EP status on CHD risk factors after soya intervention; –, adverse effect of EP status on risk factors of CHD after soya intervention.
Table 4. Randomised clinical trial results reporting the effect of soya isoflavone interventions and equol producer (EP) status on blood pressure and vasculature parameters*
DBP, diastolic blood pressure; MetS, metabolic syndrome; NEP, non-equol producer; FMD, flow-mediated dilation; SBP, systolic blood pressure; LDL-C, LDL-cholesterol; BP, blood pressure; MAP, mean arterial pressure; PWV, carotid to femoral pulse wave velocity; NMD, nitroglycerine-mediated endothelium-independent vasodilation; NOx, nitrate and nitrite; ET-1, endothelin-1; SAC, systemic arterial compliance; AIx, augmentation index; PFV, peak flow velocity; vWF, von Willebrand factor; IAC, isobaric arterial compliance; SVR, systemic vascular resistance; CIMT, carotid artery intima-media thickness; HDL-C, HDL-cholesterol; EFI, endothelial function index.
* Results are first stratified by the impact of EP status and then the impact of the soya isoflavone interventions on each of the lipid risk factors.
† +, Beneficial effect of soya isoflavones on risk factors of CHD; 0, negligible effect of soya isoflavones on risk factors of CHD; –, adverse effect of soya isoflavones on risk factors of CHD.
‡ +, Beneficial effect of EP status on risk factors of CHD after soya intervention; 0, negligible effect of EP status on CHD risk factors after soya intervention; –, adverse effect of EP status on risk factors of CHD after soya intervention.
Table 5. Randomised clinical trial results reporting the effect of soya isoflavone interventions and equol producer (EP) status on inflammation and DNA damage parameters*
CRP, C-reactive protein; sICAM-1, soluble intercellular adhesion molecule-1; MetS, metabolic syndrome; NEP, non-equol producer; hsCRP, high-sensitivity C-reactive protein; MDA, malondialdehyde; P-selectin, platelet selectin; Hcy, homocysteine; E-selectin, endothelial selectin; VCAM-1, vascular cell adhesion molecule 1; ICAM-1, intracellular adhesion molecule-1; MCP-1, monocyte chemoattractant protein-1; AP-site, apurinic/apyrimidinic site; TUNEL, terminal deoxynucleotidyl transferase dUTP nick end labelling; ADMA, asymmetric dimethylarginine.
* Results are first stratified by the impact of EP status and then the impact of the soya isoflavone interventions on each of the lipid risk factors.
† +, Beneficial effect of soya isoflavones on risk factors of CHD; 0, negligible effect of soya isoflavones on risk factors of CHD; –, adverse effect of soya isoflavones on risk factors of CHD.
‡ +, Beneficial effect of EP status on risk factors of CHD after soya intervention; 0, negligible effect of EP status on CHD risk factors after soya intervention; –, adverse effect of EP status on risk factors of CHD after soya intervention.
Table 6. Randomised controlled trial results reporting the effect of soya isoflavone interventions and equol producer (EP) status on glucose and insulin parameters*
IGF, insulin-like growth factor; IGF-BP1, insulin-like growth factor binding protein-1; SHBG, sex hormone binding globulin; IGF-BP3, insulin-like growth factor binding protein-3.
* Results are first stratified by the impact of EP status and then the impact of the soya isoflavone interventions on each of the lipid risk factors.
† +, Beneficial effect of soya isoflavones on risk factors of CHD; 0, negligible effect of soya isoflavones on risk factors of CHD; –, adverse effect of soya isoflavones on risk factors of CHD.
‡ +, Beneficial effect of EP status on risk factors of CHD after soya intervention; 0, negligible effect of EP status on CHD risk factors after soya intervention; –, adverse effect of EP status on risk factors of CHD after soya intervention.
Table 7. Randomised controlled trial results reporting the effect of soya isoflavone interventions and equol producer (EP) status on body composition variables*
BW, body weight.
* Results are first stratified by the impact of EP status and then the impact of the soya isoflavone interventions on each of the lipid risk factors.
† +, Beneficial effect of soya isoflavones on risk factors of CHD; 0, negligible effect of soya isoflavones on risk factors of CHD; –, adverse effect of soya isoflavones on risk factors of CHD.
‡ +, Beneficial effect of EP status on risk factors of CHD after soya intervention; 0, negligible effect of EP status on CHD risk factors after soya intervention; –, adverse effect of EP status on risk factors of CHD after soya intervention.
Table 8. Results of the randomised clinical trials examining the effect of soya isoflavone interventions on the risk factors for CHD in equol producers (EP) only
LDL-C, LDL-cholesterol; HDL-C, HDL-cholesterol; hsCRP, high-sensitivity C-reactive protein; TC, total cholesterol; CIMT, carotid artery intima-media thickness; DBP, diastolic blood pressure; SBP, systolic blood pressure; MAP, mean arterial pressure; FMD, flow-mediated dilation; BW, body weight; NEP, non-equol producers; TLR4, Toll-like receptor 4; TIRAP, toll–interleukin 1 receptor domain-containing adaptor protein.
* +, Beneficial effect of EP status on risk factors of CHD after soya intervention; 0, negligible effect of EP status on CHD risk factors after soya intervention; –, adverse effect of EP status on risk factors of CHD after soya intervention.
Forty studies found no association between soya isoflavones and risk factors for CHD compared with placebo. Of these, equol producer status significantly improved risk factors for CHD in seven studies (including total cholesterol, LDL-C, TAG, apoA-I, apoB, lipoprotein (a), blood pressure, diastolic blood pressure, mean arterial pressure, carotid to femoral pulse wave velocity). As with the soya isoflavone intervention, equol producer status was insignificant in thirty-two studies and was adverse in one study.
Three studies found that soya isoflavones had a negative effect on the risk factors of CHD. Of these, equol producer status was negligible in two studies and magnified the adverse outcomes of the soya isoflavone intervention in one study (isoprostane excretion). Equol producer status was also associated with the adverse outcome of an increase in insulin-like growth factor binding protein-3.
Five studies were comprised of participants who were all equol producers (Table 8); two of the studies found statistically significant beneficial effects of the isoflavone interventions on risk factors of CHD (including LDL-C, high-sensitivity C-reactive protein, TAG, inflammatory gene expression) while four studies observed negligible effects.
The RCT varied in quality, with the overall scores provided in Table 1 and the ratings summarised in Supplementary Table S3. Failure to report sample size calculations, details on the randomisation and allocation concealment procedures, and lack of intention-to-treat analyses or other suitable statistical method of dealing with participant drop-out were the most frequent flaws. Six RCT were given a ‘good’ rating, twenty were given a ‘fair’ rating and sixteen were given a ‘poor’ rating.
The heterogeneity of the studies in terms of populations, treatment regimens, intended duration, and outcomes prevented us from quantitatively synthesising the evidence in the form of a meta-analysis. Besides, the total number of participants included in all forty-two of the studies together was 3796, which, along with varying interventions and populations, probably provides insufficient statistical power to quantitatively measure the effect of dietary interventions. Further, most of these forty-two studies were small and had fewer than fifty participants, and only eighteen out of the forty-two studies qualified to be ‘fair’ or ‘good’ quality. The six ‘good’-quality papers (Hodis et al. ( Reference Hodis, Mack and Kono 14 ); Liu et al. ( Reference Liu, Ho and Chen 39 – Reference Liu, Ho and Chen 41 ); van der Velpen et al. ( Reference van der Velpen, Geelen and Hollman 53 , Reference van der Velpen, Geelen and Schouten 54 )) come from three different trials – while the Hodis study examined carotid artery intima-media thickness progression among equol producers and non-producers, Liu et al. and van der Velpen et al. examined their intervention only among equol producers. Liu et al. examined the effect of soya on risk factors such as lipid markers( Reference Liu, Ho and Chen 39 – Reference Liu, Ho and Chen 41 ), while van der Velpen et al. ( Reference van der Velpen, Geelen and Hollman 53 , Reference van der Velpen, Geelen and Schouten 54 ) examined the effect of soya on the expression of inflammatory genes. Given these varying outcomes, we have chosen to not perform a meta-analysis in our present review.
Discussion
Though the overall effect of equol producer status during a dietary soya intervention on risk factors of CHD is inconclusive, we found evidence of a favourable effect of equol producer status in fourteen of the forty-two studies( Reference Acharjee, Zhou and Elajami 27 , Reference Curtis, Potter and Kroon 30 , Reference Hall, Vafeiadou and Hallund 35 , Reference Liu, Ho and Chen 39 , Reference Mangano, Hutchins-Wiese and Kenny 42 , Reference Pusparini and Hidayat 45 , Reference Törmälä, Appt and Clarkson 50 , Reference van der Velpen, Geelen and Schouten 54 – Reference Clerici, Setchell and Battezzati 56 , Reference Meyer, Larkin and Owen 59 , Reference Pipe, Gobert and Capes 61 , Reference Wong, Kendall and Marchie 66 , Reference McVeigh, Dillingham and Lampe 67 ) regardless of the success of the soya intervention. Equol production was associated with positive changes in cholesterol( Reference Hall, Vafeiadou and Hallund 35 , Reference Liu, Ho and Chen 39 , Reference Mangano, Hutchins-Wiese and Kenny 42 , Reference Clerici, Setchell and Battezzati 56 , Reference Meyer, Larkin and Owen 59 , Reference Pipe, Gobert and Capes 61 , Reference Wong, Kendall and Marchie 66 , Reference McVeigh, Dillingham and Lampe 67 ) and other lipid variables( Reference Acharjee, Zhou and Elajami 27 , Reference Hall, Vafeiadou and Hallund 35 , Reference Liu, Ho and Chen 39 , Reference Clerici, Setchell and Battezzati 56 , Reference Meyer, Larkin and Owen 59 , Reference Pipe, Gobert and Capes 61 , Reference Wong, Kendall and Marchie 66 ), blood pressure measurements( Reference Acharjee, Zhou and Elajami 27 , Reference Curtis, Potter and Kroon 30 , Reference Welty, Lee and Lew 55 , Reference Clerici, Setchell and Battezzati 56 ) and inflammatory markers( Reference Acharjee, Zhou and Elajami 27 , Reference Liu, Ho and Chen 39 , Reference Pusparini and Hidayat 45 , Reference Törmälä, Appt and Clarkson 50 , Reference van der Velpen, Geelen and Schouten 54 , Reference Clerici, Setchell and Battezzati 56 ). The effect of equol producer status was insignificant on CHD risk factors in forty studies( Reference Hodis, Mack and Kono 14 , Reference Acharjee, Zhou and Elajami 27 – Reference Reimann, Dierkes and Carlsohn 46 , Reference Steinberg, Guthrie and Villablanca 48 – Reference McVeigh, Dillingham and Lampe 67 ) and adverse in two studies( Reference Campbell, Woodside and Honour 29 , Reference Sen, Morimoto and Heak 47 ). We did not find consistent evidence of equol production affecting specific risk factors for CHD. The heterogeneity of the CHD risk factors analysed, sample size, study designs and quality, and definition of equol producers prevented quantitative synthesis of the results.
The majority of the studies in the present review retrospectively categorised study participants by equol producer status and conducted a secondary analysis of the effect of equol on the risk factors for CHD. Therefore, these RCT were very unlikely to be sufficiently powered to detect a difference in CHD risk factors between equol producers and non-equol producers. We identified ten studies with study designs that included enrolment criteria based on equol producer status( Reference Badeau, Jauhiainen and Metso 28 , Reference Liu, Ho and Chen 39 – Reference Liu, Ho and Chen 41 , Reference Törmälä, Appt and Clarkson 49 – Reference van der Velpen, Geelen and Schouten 54 ). Of these, three found equol producer status improved several CHD risk factors (LDL-C, LDL-C:HDL-C, TAG, platelet-selectin and inflammatory gene expression) after the soya intervention( Reference Liu, Ho and Chen 39 , Reference Törmälä, Appt and Clarkson 50 , Reference van der Velpen, Geelen and Schouten 54 ) while the remaining associations measured in the RCT were negligible.
There are numerous differences in the experimental design of the RCT that could explain the inconsistency in the outcomes. The isoflavone dose ranged in both quantity and consistency between RCT. In particular, the amount of daidzein in the intervention formulations, which indicates the magnitude of equol that could be metabolised from daidzein and bioavailable in equol producers, largely varied between studies. Additionally, the duration and frequency of exposure to the intervention were inconsistent. Curtis et al. ( Reference Curtis, Potter and Kroon 30 ) found that improvements in blood pressure, mean arterial pressure, and pulse wave velocity measures in equol producers were seen after 1 year but not at 6 months, suggesting that long-term exposure to isoflavones may be more beneficial.
The criteria used to define equol producers differed across the RCT included in our review, with variability in the biological samples used to measure equol, the concentration cut-offs selected to distinguish equol producers from non-equol producers, and the analytical methods used to measure equol. Setchell & Cole( Reference Setchell and Cole 68 ) proposed classifying equol producers by a threshold log10-transformed ratio of S-(-)equol, a diastereoisomer of equol produced by the intestinal bacteria in humans, to its precursor daidzein of −1·75 in urine after a 3 d soya isoflavone challenge. This accounts for inconsistency in the technical measurements of equol and avoids classifying equol producers based on absolute measurements of equol, which exhibit greater variability( Reference Setchell and Cole 68 ). Nine studies used this approach( Reference Liu, Ho and Chen 39 – Reference van der Velpen, Geelen and Schouten 54 , Reference Clerici, Setchell and Battezzati 56 , Reference Qin, Shu and Zeng 62 , Reference Thorp, Howe and Mori 64 , Reference Wong, Kendall and Marchie 66 ), with four finding a beneficial effect of equol producer status on risk factors of CHD( Reference Liu, Ho and Chen 39 , Reference van der Velpen, Geelen and Schouten 54 , Reference Clerici, Setchell and Battezzati 56 , Reference Wong, Kendall and Marchie 66 ) and eight finding a negligible effect( Reference Liu, Ho and Chen 39 – Reference Liu, Ho and Chen 41 , Reference van der Velpen, Geelen and Hollman 53 , Reference Clerici, Setchell and Battezzati 56 , Reference Qin, Shu and Zeng 62 , Reference Thorp, Howe and Mori 64 , Reference Wong, Kendall and Marchie 66 ).
Further complicating the interpretation of the data are the potential sex differences in the metabolism of soya( Reference Ahn-Jarvis, Clinton and Riedl 69 ), which could affect the bioavailability of isoflavone metabolites between males and females. In a meta-analysis examining the effects of soya isoflavones on lipids, subjects with hypercholesterolaemia had greater reductions in men than in women( Reference Zhan and Ho 12 ). While there were studies of mixed sex (n 11) or of only males (n 1), the present review consisted primarily of female-only RCT, which may have masked the effects of equol producer status on the outcome measurements. Nestel et al. ( Reference Nestel, Cehun and Chronopoulos 60 ) found that LDL-C was significantly reduced after supplementation with biochanin (a precursor of genistein) compared with placebo (P = 0·026); when results were stratified based on sex, males showed a significant reduction in median LDL-C levels of 9·5 % while females had no measurable difference. Equol producer status did not further reduce LDL-C, which the authors speculated was due to the small sample size of fifteen equol producers, with seven included in the biochanin intervention group( Reference Nestel, Cehun and Chronopoulos 60 ).
The source of soya may also contribute to the variability in its effectiveness. The type of processing used for soya products during production can affect the isoflavone content( Reference Sacks, Lichtenstein and Van Horn 13 ) and modify other components of soya( Reference Gianazza, Eberini and Arnoldi 70 ). Additionally, soya protein isolate primarily contains isoflavone glucosides while fermented soya foods contain isoflavones mainly in the aglycone form( Reference Setchell and Clerici 15 , Reference Clerici, Nardi and Battezzati 71 ). Isoflavone aglycones are absorbed more efficiently than isoflavone glucosides in humans and may therefore be more effective in CHD prevention( Reference Pan, Ikeda and Takebe 72 ). Daidzein in the aglycone form is also more readily converted to equol( Reference Setchell and Clerici 15 ). Clerici et al. ( Reference Clerici, Setchell and Battezzati 56 ) found that pasta enriched in isoflavone aglycones significantly reduced total cholesterol, LDL-C, high-sensitivity C-reactive protein, and arterial stiffness compared with placebo in study participants, with effects more pronounced in equol producers. Of the fourteen RCT that found a positive association between equol producer status and CHD risk factors, seven used interventions of foods and milk enriched with soya( Reference Acharjee, Zhou and Elajami 27 , Reference Curtis, Potter and Kroon 30 , Reference Hall, Vafeiadou and Hallund 35 , Reference Welty, Lee and Lew 55 , Reference Clerici, Setchell and Battezzati 56 , Reference Meyer, Larkin and Owen 59 , Reference Wong, Kendall and Marchie 66 ).
Furthermore, baseline age and the health status of the participants may contribute to variability in the outcome measurements. Oestrogen receptor β has been found to be enhanced in extracted arteries from postmenopausal CHD patients compared with normal subjects, with enhanced dilation in response to isoflavones( Reference Cruz, Agewall and Schenck-Gustafsson 73 ). Hodis et al. found that isoflavone supplementation failed to prevent the progression of subclinical atherosclerosis in healthy postmenopausal women overall; a subanalysis indicated, however, that healthy women within 5 years of becoming postmenopausal had a significantly reduced mean carotid artery intima-media thickness progression rate of 68 % compared with placebo( Reference Balk, Chung and Chew 74 ). Previous meta-analyses have also found lipid variables to be more positively affected by soya interventions in hypercholesterolaemic patients than in healthy subjects( Reference Zhan and Ho 12 , Reference Anderson, Johnstone and Cook-Newell 75 ). We identified thirty-five RCT that only used postmenopausal women; all of the studies that found a favourable association of equol producer status on risk factors of CHD had postmenopausal participants. There were a relatively equal number of RCT using healthy participants (n 20) v. participants with underlying health issues or a history of illness (n 22); of the fourteen studies that found equol producer status to improve risk factors for CHD, five had healthy participants( Reference Hall, Vafeiadou and Hallund 35 , Reference Mangano, Hutchins-Wiese and Kenny 42 , Reference Pusparini and Hidayat 45 , Reference van der Velpen, Geelen and Schouten 54 , Reference McVeigh, Dillingham and Lampe 67 ) while nine had participants with underlying health issues related to CHD( Reference Acharjee, Zhou and Elajami 27 , Reference Curtis, Potter and Kroon 30 , Reference Liu, Ho and Chen 39 , Reference Törmälä, Appt and Clarkson 50 , Reference Welty, Lee and Lew 55 , Reference Clerici, Setchell and Battezzati 56 , Reference Meyer, Larkin and Owen 59 , Reference Pipe, Gobert and Capes 61 , Reference Wong, Kendall and Marchie 66 ).
In the present systematic review, electronic databases were extensively searched following our defined set of guidelines and used to extract relevant data. Our results may imply that equol is beneficial on cardiovascular health, yet the interpretation is limited largely because of the secondary analysis of equol producers in RCT of dietary sources of isoflavones. Recently, equol itself has become available as a dietary supplement. Orally administered equol has greater plasma accumulation than other dietary sources of isoflavones( Reference Legette, Prasain and King 76 ) and has the potential for enhanced therapeutic effects due to its more potent antioxidant properties and bioactivity among all isoflavones. In fact, one RCT of equol on risk factors of CHD has been conducted. Usui et al. ( Reference Usui, Tochiya and Sasaki 77 ) found a statistically significant improvement in LDL-C, glycated HbA1c levels, and cardio-ankle vascular index scores, a measure of vascular stiffness, in overweight and obese patients after dietary equol supplementation, particularly for non-equol producers. This study is limited by its small sample size and short duration of the intervention. We recommend additional RCT of equol itself as an intervention to directly assess its effects on CHD risk factors and potentially CHD.
Supplementary material
The supplementary material for this article can be found at http://dx.doi.org/10.1017/jns.2016.18
Acknowledgements
We thank Barb Folb for her expertise and guidance in designing the systematic review.
The present review was supported by the National Institutes of Health (R01 HL068200).
The authors’ responsibilities were as follows: A. S. designed the study; R. L. B. conducted the research; A. S., A. V. and R. L. B. analysed the data; A. S., V. A., A. V., R. W. E., Y. M., K. M., T. U. and R. L. B. drafted the manuscript; A. S., V. A. and R. L. B. had primary responsibility for the final content of the manuscript; and all authors read and approved the final manuscript.
None of the authors reported a conflict of interest related to the present review.
Open access
