Scholarly article on topic 'Daidzein and genistein fail to improve glycemic control and insulin sensitivity in Chinese women with impaired glucose regulation: A double-blind, randomized, placebo-controlled trial'

Daidzein and genistein fail to improve glycemic control and insulin sensitivity in Chinese women with impaired glucose regulation: A double-blind, randomized, placebo-controlled trial Academic research paper on "Health sciences"

Share paper
Academic journal
Molecular Nutrition & Food Research
OECD Field of science

Academic research paper on topic "Daidzein and genistein fail to improve glycemic control and insulin sensitivity in Chinese women with impaired glucose regulation: A double-blind, randomized, placebo-controlled trial"

Research Article

Daidzein and genistein fail to improve glycemic control and insulin sensitivity in Chinese women with impaired glucose regulation: A double-blind, randomized, placebo-controlled trial

Yan-Bin Ye1,2,3, Ai-Ling Chen3, Wei Lu1, Shu-Yu Zhuo1, Juan Liu3, Jian-Hua Guan4, Wan-Ping Deng3, Shi Fang1, Yan-Bing Li3* and Yu-Ming Chen2

1 Department of Nutrition, The First Affiliated Hospital of Sun Yat-sen University, Guangzhou, P R. China

2 Guangdong Provincial Key Laboratory of Food, Nutrition and Health, School of Public Health, Sun Yat-sen University, Guangzhou, P R. China

3 Department of Endocrinology, The First Affiliated Hospital of Sun Yat-sen University, Guangzhou, P R. China

4 Department of Endocrinology, The Clifford Hospital of Traditional Chinese Medicine, University of Guangzhou, Guangzhou, P. R. China

Scope: This randomized, double-blind, and placebo-controlled trial evaluated the effect of isolated daidzein and genistein on glycemic control and insulin sensitivity in 165 Chinese women aged 30-70 with impaired glucose regulation (IGR).

Methods and results: Participants were randomly assigned to one of three groups with a daily dose of 10 g of soy protein plus (i) no addition, (ii) 50 mg of daidzein, or (iii) 50 mg of genistein for 24 wk. Fasting glucose (FG), insulin, and glycosylated hemoglobin (HbA1c), and glucose concentrations at 30, 60, 120, and 180 min and insulin concentrations at 30, 60, and 120 min after an oral 75-g glucose tolerance test were assessed at baseline and at 12 and 24 wk postintervention. a total of 158 and 151 subjects completed the measures at wk 12 and 24, respectively. There were no significant differences in the changes (%) of FG and the 2-h glucose, HbA1c, fasting, and 2-h insulin or the area under the curve of glucose and insulin between the three treatment groups at wk 12 or 24 (all p > 0.05).

Conclusion: Neither isolated daidzein nor genistein has a significant effect on glycemic control and insulin sensitivity in Chinese women with IGR over a 6-month supplementation period.


Daidzein / Genistein / Glycemic control / Insulin sensitivity / Soy isoflavones

Received: June 5, 2014 Revised: October 21, 2014 Accepted: October 22, 2014

Additional supporting information may be found in the online version of this article at the publisher's web-site

1 Introduction

Type 2 diabetes is a well-known, serious, and chronic disease with a high prevalence globally. The International

Correspondence: Professor Yu-Ming Chen, Department of Medical Statistics and Epidemiology, School of Public Health, Sun Yat-sen University, 74# Zhongshan Road 2, Guangzhou, 510080, P R. China

E-mail: Fax: +86-20-87330446

Abbreviations: FG, fasting glucose; FIns, fasting insulin; HbAic, glycosylated hemoglobin; IGR, impaired glucose regulation; OGTT, oral glucosetolerancetest; PG, postprandial glucose; RCT, randomized controlled trial

Diabetes Federal (IDF) has estimated that the number of cases of diabetes will increase from 366 million in 2011 to 552 million by 2030 [1]. Currently, the prevalence of diabetes and prediabetes in Chinese adults is 9.7 and 15.5%, respectively [2]. Given the huge economic burden ofdiabetes worldwide, effective prevention and control strategies are essential [3].

Soy foods are part of the traditional cuisine in Asia. Many animal and human studies have examined the association

* Additional corresponding author: Dr. Yan-bing Li, E-mail:

Clinical Trial Registry: NCT 00951912.

of soy foods or isoflavones with type 2 diabetes. A few animal studies have suggested that soy protein and isoflavones may improve blood glucose, insulin sensitivity, and insulin requirements [4,5] by upregulating the expression of PPAR-a and -7 genes [6], thus increasing the affinity of insulin receptor [7]. Human studies have generated inconsistent results. A cohort study in Shanghai found that the highest intake of soybean (32 g/day versus the lowest 2.8 g/day) significantly reduced the risk of type 2 diabetes by 47% (95% CI: 38-55%) in middle-aged Chinese women [8]. Similar results were observed in another cohort study in Singapore [9]. However, such a favorable association was not observed in two studies among Caucasians, Japanese, and Natives in Hawaii [10] and Japanese in Japan [11]. Inconsistent null results were also obtained in human intervention studies ofmixed isoflavones [12,13], but the former also reported the favorable effects of genistein on glucose and insulin. Some small-scale human studies [14-17], but not all [18,19], found that soy protein or soy isoflavones might improve blood glucose and insulin sensitivity. Moreover, some studies showed that the aglycone forms of isoflavones had a faster and more efficient absorption than the glycoside and mixed forms [20, 21]. Weaver et al. reported that genistein and daidzein (the two main components of soy isoflavones) may be resistant to each other [22], and previous intervention studies suggested that isolated genistein or daidzein was more effective in the improvement of bone health than mixed isoflavones [22-24]. Therefore, the effect of soy foods or isoflavones on diabetes remains inconclusive, and further studies are needed to determine the effects of isolated genistein and daidzein on the risk of diabetes or glycemic control in humans.

Genistein is the most abundant isoflavone in soy. Previous studies found that genistein might improve glycemic control by tyrosine kinase inhibitory action [25], improving insulin sensitivity [26], increasing insulin secretion [5], and preserving pancreatic p-cell function [27]. Four randomized controlled trials (RCTs) reported that isolated genistein may improve fasting glucose (FG) and insulin resistance in postmenopausal women with normal blood glucose [14, 28-30], but not in diabetic patients. The hypoglycemic effects of genistein in persons with impaired glucose regulation (IGR) are unclear. Daidzein is the second most abundant isoflavone. Fewer studies have assessed its effects compared to genistein. Daidzein might play a role in health mainly through its intestinal metabolite, equol [31], which has the greatest affinity for the estrogen receptor-p. Equol exists in two enantiomeric forms, R-(+)equol and S-(-)equol, and the latter is the natural diastereoisomer produced by intestinal bacteria in the intestines ofhumans [32]. Some studies have indicated that S-equol is the real effective component in bone health and cardiovascular diseases [31], but no human study has yet tested the effects of either equol on glycemic control in diabetic persons and no study has compared the effects of genistein and daidzein on glycemic control.

To address these issues, we undertook a 24-wk RCT to examine and compare the effects of two isolated soy isoflavones

(genistein and daidzein) on glycemic control and insulin sensitivity in women with IGR.

2 Materials and methods

2.1 Participants and study design

A total of 165 Chinese women were recruited for this randomized, double-blinded, and placebo-controlled trial from local communities, participants attending regular health examinations, or relevant outpatients between March 2009 and November 2009 in Guangzhou, China. The participants were required to be aged 30-70, with a FG concentration of 5.6-7.0 mmol/L, a 2-h postprandial glucose (PG) concentration of 7.8-11.0 mmol/L, or newly diagnosed diabetes not requiring medication treatment according to the doctor's recommendations. We excluded those with confirmed coronary heart disease; stroke; thyroid disease; liver dysfunction; lung, kidney, or gastrointestinal tract diseases, current use of hypoglycemic, lipid-lowering or weight-reduction agents, or hormone replacement therapy in the past 8 wk; diabetic complications; allergies to soy; or long-term antibiotic use. A stepwise screening strategy was used to identify potential eligible subjects. Telephone or in-person interviews were used to screen participants with high risk of IGR or diabetes, including being overweight or obese, a family history of diabetes, a history of gestational diabetes, IGR, hypertension, or dyslipidemia. Their final eligibility was further confirmed by plasma FG and PG of a 75-g oral glucose tolerance test (OGTT). The ethics committee of the First Affiliated Hospital of Sun Yat-sen University approved the study (no. FAH 2008016), and all participants signed a written informed consent form before final enrollment.

2.2 Randomization and blinding

A block randomization method was adopted to assign the participants to three groups. In each block, 15 serial numbers were classified into tertiles according to the sequence of computer-generated random numbers. Another three random numbers were used to determine their intervention groups. The three types of apparent identical supplements were labeled with one corresponding serial number from 1 to 165 according to the randomization results. The corresponding interventions were assigned to the participants according to the order offinal enrollment after the completion ofa 2-wk run-in period. All of the participants and investigators who might contact participants or be involved in data collection or data analysis were blinded to the group status until the completion of the data analyses. To evaluate the blinding efficacy, subjects were asked which group they thought they were in at the final visit. The proportions of correct estimations for the three groups (Placebo, Daidzein, and Genistein)

were 12.8, 24, and 25.9, respectively (p = 0.228), implying that the blinding was successful.

2.3 Intervention

The eligible subjects received a placebo for a 2-wk run-in study, and participants with good compliance (consumed > 90% supplements) and without side effects were assigned to one of the three study groups with a daily dose of 10 g of soy protein isolated as a substrate and supplemented with (i) nothing (Placebo group), (ii) 50 mg of daidzein (Daidzein group), and (iii) 50 mg of genistein (Genistein). The soy protein was isolated with a very low level of isoflavones (<200 mg/kg) and was purchased from Solae Ltd. (St Louis, MO, USA). The aglycone type of genistein (98.23%) and daidzein (98.12%) were purchased from Xi'an Chongxin Natural Additive Limited Company (Xi'an, China). Soy beans contain 1.1-1.5 mg of total isoflavones per gram, which are naturally present as ^-glycosides of genistin, daidzin, and glycitin, representing 50-55%, 40-45%, and 5-10% of the total isoflavones content, respectively. Therefore, the intervention dose of 50 mg of aglycone isoflavones equals more than 100 mg of total isoflavones and would require the consumption of approximately 70 g of soybean or equivalent soy products.

The supplements were prepared as an isocaloric powder with similar color and flavor. Each dose was filled into an identical packet. The uniformity ofthe intervention dose and other nutrients were verified by testing random samples before and after packaging. The supplements were delivered to participants at wk 1, 12, and 20, and the remaining samples were recorded and counted at the following visits. The participants were told to take one dose as a part of breakfast, to maintain their usual lifestyles including diet and physical activities, and to avoid any supplements known to affect glucose or lipid metabolism, such as dietary fiber, fish oil, or any other supplements containing isoflavones, during the intervention period. Because of the low background soy food consumption (12-13 mg/day isoflavones), and in consideration of the ethics, we did not limit soy food consumption during the study.

2.4 Data collection

At baseline and wk 12 and 24, overnight fasting (>10 h) venous blood samples were collected to determine FG and insulin, and blood samples were also collected at 30,60,120, and 180 min after the OGTT to assess PG and insulin. Blood containing EDTA was used for glycosylated hemoglobin (HbA1c) within 180 days. The baseline and follow-up samples for each individual were measured in the same batch. Twenty-four-hour urine was collected for the determination of urinary isoflavones and isoflavones metabolites at wk 12.

Individual general information, including sociodemo-graphic data, medical history, dietary habits (a validity food

frequency questionnaire), and physical activities, was collected at the baseline via face-to-face interviews using a structured questionnaire. Dietary intake was assessed using 3-day diet records completed by the participants prior to each follow-up visit at wk 12 and 24. Approximately 30 min of training was given to instruct participants in food consuming records, and these records were checked during face-to-face interviews. Dietary nutrients and isoflavones were calculated according to the 2004 China Food Composition Table [33]. Physical activity was assessed on the basis of a validated questionnaire at each follow-up visit.

2.5 Anthropometric and biochemical measurements

2.5.1 Anthropometric measurements

Body height and weight, waist circumference, sitting blood pressure, and body fat levels were measured according to the standard protocols at baseline and wk 12 and 24. The total body fat (%) was assessed with a bioelectrical impedance analyzer (BIA, TBF-410-GS Tanita Body Composition Analyzer, Japan).

2.5.2 Outcome measurements

We used a glucose oxidized method to determine the glucose concentration (AEROSET analyzer and kit, Abbott, USA). Serum insulin was measured using a chemiluminescent microparticle immunoassay (ARCHITECT automatic analyzer and kit, Abbott). HbA1c was measured by an HPLC (VARIANT™Ilautomatic analyzer for HbA1c, Bio-Rad Co. Ltd., USA). Intra- and interassay CVs were 1.8 and 3.6% for glucose, 4.0 and 5.8% for insulin, and 2.0 and 2.8% for HbA1c, respectively. All samples were analyzed in parallel with the respective quality controls provided by the supplier on the same day and in the same assay.

Insulin sensitivity indices were calculated as: (1) HOMA-IR = (fasting insulin (FIns, ^U/L) x FG (mmol/L))/22.5. (2) The insulin sensitivity index (ISI) = 10 000/SQRT ((FG x FIns) x (GOGTTmean X IOGTTmean)). (3) The P<ell of HOMA (HOMA-p,%) = FIns x 20/(FG - 3.5). (4) The quantitative insulin sensitivity check index (QUICKI) = 1/(log FIns + log FG). The area under the curve for glucose and insulin was calculated from the OGTT data based on the trapezoid method, and the glucose and insulin maximum concentrations (Cmax) and time to Cmax (Tmax) were identified from the OGTT data for glucose and insulin [19].

2.5.3 Urinary isoflavones measurement

The participant urinary isoflavones and metabolites levels were analyzed using HPLC (Waters e2695 chromatography, Waters Co. Ltd., USA) [34]. The intra- and interassay CVs

of the total isoflavones measurement were 0.61 and 2.78%, respectively.

2.6 Statistical analyses

The sample-size for this study was estimated according to a change of FG of 8 mg/dL (percent change was 7%) and an SD for the change of 12.1 mg/dL reported in Hong Kong Chinese women [35]. A sample of 36 participants would have a power level of 80% to detect the difference in FG changes between the treatment and placebo groups at the a level of 0.05. Given a 20% withdrawal rate, 43 subjects were required in each group. We actually recruited 55-57 participants in each group. The actual power was 98.2% to detect the 7% change in FG according to the current SD (7.46 and 7.10) and study size.

All of the statistical analyses were performed with SPSS 16.0 software (SPSS Inc., Chicago, IL, USA). The data with skewed distributions were corrected by log-transformation before analysis and reported as the geometric mean. The mean comparisons of continuous variables were evaluated with one-way ANOVA. A chi-square test was used to compare the categorical data. The outcome measures include indices of glucose and insulin and insulin resistance, and their percentage changes were calculated as follows: (follow-up value -baseline value)/baseline value x 100%. ANOVA for repeated measures was used to compare the changes in each outcome variable among the intervention arms over time (0, 12, and 24 wk). ANOVA and analysis of covariance (ANCOVA) with adjustment for the relevant baseline variable and the change of body weight were conducted to compare the outcome measures at 12 and 24 wk and their percentage changes, respectively. The Bonferroni test was used for post hoc multiple comparisons among the three groups. We also used interaction analyses to examine whether the effect of soy isoflavones supplementation differed across various subgroups classified according to the baseline concentrations of FG and PG, and, in the daidzein group, equol producer status. The equol producer was determined as the ratio of equol/daidzein > 0.018, with a daidzein threshold >2 nmol/mg creatinine [36].

Intention-to-treat analyses were undertaken regardless of compliance by including all randomized 165 participants. The last value was carried forward for missing data at follow-up. The secondary analyses was composed of a per protocol analysis that only included 151 participants with good compliance, which was defined as consuming >80% supplements and completing all assessments. A p-value of 0.05 or less was considered statistically significant.

3 Results

3.1 Study flow, compliance, and adverse effects

Among 165 women, 158 and 151 participants attended the visits at wk 12 and 24, respectively. Seven, five, and two

participants dropped out of the Placebo, Daidzein, and Genistein groups, respectively, for various reasons (Fig. 1). Approximately 95% of the participants consumed >97% of the supplements. The 24-h urinary excretion of daidzein (and equol) was much higher in the Daidzein group, while genistein was much greater in the Genistein groups than in the other two groups at wk 12 (all p < 0.05). (Supporting Information Table 3)

No severe adverse events were reported during the study. We documented 30 cases of mild or moderate adverse events, including 26 cases of mild gastrointestinal discomfort (e.g. bloating, flatulence, constipation, diarrhea, and thirst), 2 cases of mild skin sensibility, 1 case of hyperplasia of the mammary glands, and 1 case of albuminuria. There were no significant differences in the incidence of these adverse events among the three groups.

3.2 Baseline characteristics, diet, and physical activity at follow-up

The participants' baseline characteristics of age, work status, education, body weight, BMI, WC, body fat, dietary intake, and physical activity were comparable in the three groups (p > 0.05). The participants' body weight, BMI, WC, body fat, dietary intake, and physical activity decreased over the follow-up period, and the changes of these markers were similar in the three groups (Table 1 and Supporting Information Table 2). There were no significant differences between the missing and remaining participants (Supporting Information Table 1).

3.3 Effects of soy isoflavones on glycemic control and insulin resistance

No significant differences in the biomarkers of glycemic control and insulin resistance were found among the three groups at baseline and at wk 12 and 24 in IGR women without any drug treatments. Moreover, we did not find any significant differences in the changes or percentage changes in these outcomes among the three arms. The mean percent changes (%) over the 24-wk intervention in the Placebo, Daidzein, and Genistein groups, respectively, were as follows: FG: -1.80, -1.0, and -4.89; 120-m PG: -4.9, -3.13, and -5.59; HbAk: -1.38, -1.37, and 1.30; FIns: 25.27, 21.41, and 6.06; HOMA-IR: 22.83, 21.60, and-0.22; ISWsud*: 20.46,9.28, and 27.61; and QUICKI: 2.32, 0.18, and 6.29 (all p > 0.05). The results of ANOVA for repeated analysis also showed no significant interactions between study time and the interventions in these indices with or without adjusting for the body weight changes (Tables 2 and 3). Similar null effects were observed when adjusting for baseline values (ANCOVA) and in the per protocol analysis, including only the 151 valid completers (data not shown).

924 individuals screened

759 excluded

694 no blood tests

53 normal glucose

12 dropout after run-in 12

165 randomly assigned

54 received placebo

55 received daidzein 50 mg

56 received genistein 50 mg

4 withdrawal 2 noncompliance 1 physician decision 1 other disease

50 Attended 12 week visit 54 Attended 12 week visit

1 withdrawal 1 adverse event

3 withdrawal 2 noncompliance 1 other disease

47 Attended 24 week visit

4 withdrawal 2 noncompliance 1 other disease 1 lost to follow-up

54 in the ITT analysis

47 in the Per protocol analysis

2 withdrawal 1 physician decision 1 other disease

54 Attended 12 week visit

50 Attended 24 week visit 54 Attended 24 week visit

i i r

55 in the ITT analysis 50 in the Per protocol analysis 56 in the ITT analysis 54 in the Per protocol analysis

Figure 1. Flowchart of the study participants .

Subgroup analyses were performed according to the status of baseline FG and PG (< or >median). All of these analyses showed a similar effect of the three interventions on the above-mentioned outcomes (p-interactions > 0.05) (Supporting Information Table 4). We also conducted stratified analyses by the equol producer status and menopause status. No significant effect was found in any subgroup (Supporting Information Tables 5 and 6).

4 Discussion

This 24-wk trial examined and compared the effects of two isolated isoflavones (daidzein and genistein) on glucose metabolism and insulin sensitivity. We did not find any significant effect of either of the two isolated isoflavones on these outcomes in untreated diabetic patients and IGT patients. This was the first study to test the effects of isolated daidzein and to compare its effects with those of isolated genistein in IGR populations. Our findings do not support the hypothesis that isolated genistein or daidzein can improve glycemic control and insulin resistance in IGR women.

Many studies have examined the effects of isoflavones or soy products on blood glucose, including in vitro and animal [27, 37], observational [8-11, 38], and RCT [13, 14, 28-30] studies. Consistent with our findings, two RCTs found a similar null effect of soy protein containing isoflavones or isolated isoflavones on glycemic control and insulin resistance in postmenopausal women with IGR [18, 19]. The null effect was also supported by two meta-analyses of RCTs [12, 13]. In a meta-analysis of 24 RCTs, Liu et al. reported no significant mean differences in the changes in FG (-0.69, 95% CI: -1.65, 0.27) mg/dL and insulin (-0.18, 95% CI: -0.70, 0.34) ^U/mL between soy groups (protein and/or isoflavones) in all subjects with normal or abnormal blood glucose [13]. Similar results were observed in another meta-analysis in type 2 diabetes subjects [39]. However, the majority of in vitro, animal [4,5,26], and observational human studies [8,9,11,38] found a favorable association of soy foods, isolated soy protein with/without isoflavones, and isolated soy isoflavones on glycemic control, insulin resistance, or the incidence of type 2 diabetes. Two meta-analysis studies of RCTs also reported such a favorable effect in non-Asian women [12, 40]. One meta-analysis found that isolated

Table 1. The baseline and follow-up characteristics of the 165 study participants in the three treatment armsa)

Placebo (n = 54) Daidzein (n = 55) Genistein (n = 56) p Valueb)

Age (years) 56.3 ± 11.1 56.4 ± 9.9 57.0 ± 9.68 0.92

Menopausal status (n (%)) 0.75

Postmenopause 41 (75.9) 39 (70.9) 43 (76.8)

Premenopause 13(24.1) 16 (29.1) 13(23.2)

Years since menopause (years) 11.4 ± 7.3 11.2 ± 7.6 11.0 ± 6.6 0.95

Job status (n (%)) 0.73

Full-time 11 (20.4) 15 (27.3) 13(23.2)

Part-time 4(7.5) 7 (12.8) 5 (8.9)

Housewife 39 (72.2) 33 (60) 38 (67.8)

Education (n (%)) 0.93

Primary 4 (7.4) 2 (3.6) 3 (5.4)

Middle school 22 (40.8) 25 (45.5) 25 (44.6)

University 28(51.8) 28 (50.9) 28 (50)

Family history of diabetes (n (%)) 26 (48.1) 27 (49.1) 22 (39.3) 0.52

Anthropometric measurement baseline

Body weight (kg) 57.7 ± 9.7 59.4 ± 8.3 58.6 ± 9.0 0.63

BMI (kg/m2) 23.7 ± 3.5 24.5 ± 3.3 24.1 ± 3.4 0.59

Body fat (%) 31.9 ± 7.7 32.8 ± 8.5 33.0 ± 6.1 0.78

Waist circumference (cm) 82.3 ± 8.3 83.5 ± 7.0 81.7 ± 8.8 0.51

Change of follow-up and baseline

Body weight (kg) -1.3 ± 2.4 -1.3. ± 2.5 -0.7 ± 2.1 0.32

BMI (kg/m2) -0.5 ± 0.9 -0.5 ± 1.0 -0.3 ± 0.9 0.44

Body fat (%) -0.9 ±2.4 -0.6 ±2.3 -0.5 ± 2.6 0.72

Waist circumference (cm) -1.8 ± 3.5 -1.3 ± 4.2 -0.2 ± 4.1 0.11

a) Continuous and categorical variables are presented as mean ± SD and number (proportion).

b) ANOVA for continuous variables and chi-square test for categorical variables.

genistein rather than mixed isoflavones had a significant beneficial effect on FG in peri- and postmenopausal non-Asian women [12]. Another meta-analysis in 2013 found that mixed isoflavones significantly improved both FG and FIns levels in non-Asian postmenopausal women [40].

Several reasons might explain the discrepancies among the previous studies. First, positive, but not negative, results from in vitro, animal, and observational human studies [8,27,37] tended to be published instead RCTs [18,19]. Second, postmenopausal non-Asian women seemed to be more sensitive than their Asian counterparts to the effects of soy foods, soy protein, or isoflavones, possibly because of the much lower background consumption of these products in the Western diet [11,40]. However, the very low habitual consumption of soy protein (3 g/day) and isoflavones (13 mg/day) among our participants is unlikely to be the major reason for the null effect observed in this study. Third, there is a higher risk for false positives in post hoc analyses than in primary analyses [12,13,40]. Fourth, the favorable effects tended to be more significant in subjects with higher (versus normal) baseline glucose levels [18,35]. Fifth, most studies had insufficient power to detect a small effect [12,13]. Sixth, a longer supplementation or observational duration seems to identify more pronounced effects [29,30,35]. Last, many clinical trials also examined the effects of soy isoflavones or other soy products on glycemic control and insulin resistance compared to a variety of alternative controls, such as milk protein, carbohydrates, and meat protein [12,13].

Previous studies suggested that genistein might improve glucose metabolism via a variety of mechanisms, as described in the introduction. Up until this date, five RCTs have suggested that genistein was effective [14,28-30,41] in healthy women, and four of them are from the same study group at the University of Messina. Our results did not support the hypothesis that genistein may have a benefit on glycemic control in our study population. The null effect was unlikely to be due to an insufficient interventional dosage (50 mg/day), a small study size, or poor compliance. We choose the supplemental dosages of genistein and daidzein according to the findings of previous studies, many of which showed that 50 mg/day doses of genistein were effective and safe for humans [14,29]. Observational studies in Asian populations found a significant dose-dependent effect of soy food intake on diabetes within a range of 4.1 g of soy protein per day to 12.6 g of soy protein per day [38] and equal to 12-38 mg of isoflavones per day.

No RCT has yet examined the effect of isolated daidzein on glucose metabolism. One study performed a post hoc analysis of an RCT and found no significant effect of 40 and 60 mg/day doses ofdaidzein-rich isoflavones aglycones (70%) on FG levels [42]. Previous studies suggested that the daidzein's metabolite equol might play a key role in the biological effects of isoflavones. A few human studies showed that equol producers had better profiles of cardiovascular risk factors or some cancers [43]. However, we did not find any significant effects of daidzein supplementation on the endpoints related

Table 2. A comparison of the mean differences in glucose metabolism indices according to three groups among 165 women by intention-to-treat analysis8'

Placebo (n = 54) Daidzein (n = 55) Genistein (n = 56) P1b) P2b) P3b)

Fasting glucose (mmol/L) 0.29

Baseline 5.8 ± 0.9 5.7 ± 0.8 5.8 ± 0.9 0.58

Wk 12 5.7 ± 0.9 5.7 ± 1.0 5.6 ± 0.9 0.70

Wk 24 5.7 ± 1.1 5.6 ± 0.8 5.5 ± 0.9 0.69

%Change (12-0 wk) -1.5 ± 12.2 2.1 ± 18.8 -3.6 ± 2.3 0.15

%Change (24-0 wk) -1.8 ± 13.5 -1.0 ± 12.6 -4.9 ± 11.8 0.24 0.10

120-m postload glucose 0.88


Baseline 11.0 ± 3.7 11.2 ± 2.7 10.8 ± 3.4 0.73

Wk 12 9.8 ± 3.4 10.1 ± 2.8 9.2 ± 2.9 0.33

Wk 24 10.3 ± 3.7 10.7 ± 3.0 9.8 ± 3.0 0.35

%Change (12-0 wk) -9.1 ± 18.0 -9.0 ± 20.1 -11.2 ± 24.1 0.83

%Change (24-0 wk) -4.9 ± 0.7 -3.1 ± 20.6 -5.6 ± 24.2 0.83 0.63

Glucose Cmax (mmol/L) 0.20

Baseline 13.3 ± 3.1 13.0 ± 2.1 13.0 ± 3.1 0.84

Wk 24 12.5 ± 2.9 12.9 ± 2.8 12.4 ± 2.5 0.64

%Change (24-0 wk) -5.0 ± 11.3 0.2 ± 19.5 -2.6 ± 17.9 0.28 0.25

Glucose Tmax (min) 0.048

Baseline 59.2 ± 27.6 68.1 ± 28.7 62.3 ± 23.9 0.15

Wk 24 54.8 ± 23.3 67.2 ± 29.3 51.1 ± 20.7c) 0.001

%Change (24-0 wk) -1.9 ± 35.0 8.2 ± 47.9 -11.2 ± 36.0 0.06 0.04

Area under the curve of 0.58

glucose (mmol/L x

180 min)

Baseline 1874±460 1876 ± 358 1855 ± 448 0.96

Wk 24 1750 ± 483 1804 ± 436 1725 ± 391 0.63

%Change (24-0 wk) -6.2 ± 14.1 -3.3 ± 17.0 -5.2 ± 18.0 0.63 0.51

HbA1C (%) 0.38

Baseline 6.4 ± 1.1 6.4 ± 0.6 6.4 ± 0.8 0.74

Wk 12 6.4 ± 1.1 6.4 ± 0.7 6.3 ± 0.6 0.91

Wk 24 6.3 ± 0.7 6.4 ± 0.6 6.4 ± 0.8 0.62

%Change (12-0 wk) -0.1 ± 9.3 -0.7 ± 7.9 0.08 ± 7.6 0.87

%Change (24-0 wk) -1.4 ± 7.2 -1.4 ± 6.0 1.3 ± 9.3 0.11 0.14

a) All of the values presented in this table are unadjusted (mean ± SD) except that geometric mean was use for glucose and glucose 7"max; Cmax, maximum concentration; Tmax, time to Cmax; HbA1c, glycosylated hemoglobin.

b) p1: ANOVA; p2: p-interaction (time x group) by ANOVA for repeated measures; p3: ANCOVA, covariate with adjusted for body weight change at follow-up and baseline (24 wk-baseline).

c) Compared with the placebo, p < 0.05.

to glucose metabolism in equol producers, nonproducers, or the combination of the two. Recently, one study reported the favorable effects of an S-equol intervention on HbA1c and LDL in overweight women with metabolic syndromes [44]. Due to the limited study size of the equol producer sample in our study and the low transformation efficiency of equol from daidzein in humans, further studies are needed to directly examine the effect of equol on glucose metabolism.

Previous studies showed that baseline blood glucose might influence the effects of soy products on glycemic control [17, 35,45]. Ho et al. reported greater decreases of FG in subjects with higher (versus lower) baseline FG levels [35]. Other soya intervention studies have also found significant improvements in glucose, HbA1c, and insulin in diabetic adults with high levels of baseline glucose [17,45]. However, the null effect in the present study is unlikely to be due to the low baseline blood glucose because all of our subjects had an

increased FG or PG without treatment. Apart from one participant in the Genistein group who dropped out to take an antidiabetic treatment, none ofthe participants took medicine to lower glucose, lipids, or hormones. A few participants took antihypertensive drugs at the same dosage throughout the intervention period, but no study has demonstrated that these drugs could affect glucose metabolism. Therefore, the effect of genistein/daidzein could not be masked by drug treatments.

Body weight or fat, dietary intake, and physical activity might also mask the effects of the supplements. To avoid the influence of these factors, all of the participants were instructed to maintain their usual dietary and physical activity habits. Although dietary intakes of total energy, total protein, fat, soy protein, and soy isoflavones decreased during the follow-up period as a general effect of the intervention, the changes in dietary factors and physical activity were not

Table 3. A comparison of the mean differences in insulin resistance indices according to the three groups among 165 women by intention-to-treat analysis8'

Placebo (n = 54) Daidzein (n = 55) Genistein (n = 56) P1b) P2b) P3b)

Fasting insulin (pU/L) 0.44

Baseline 8.2 ± 10.5 8.2 ± 9.1 8.5 ± 5.5 0.95

Wk 24 7.9 ± 5.2 8.3 ± 6.1 7.5 ± 5.9 0.60

%Change (24-0 wk) 25.3 ± 114 21.4 ± 91.2 6.1 ± 78.3 0.49 0.42

120-m postload insulin 0.87


Baseline 54.6 ± 61.6 61.6 ± 71.0 53.1 ± 56.1 0.60

Wk 24 60.3 ± 92.3 61.3 ± 63.5 54.0 ± 47.2 0.62

%Change (24-0 wk) 26.9 ± 78.4 27.0 ± 119 51.4 ± 293 0.57 0.71

Insulin Cmax (pU/L) 0.25

Baseline 63.5 ± 93.5 70.8 ± 74.7 67.3 ± 54.7 0.84

Wk 24 72.1 ± 0.7 67.9 ± 62.8 67.3 ± 54.0 0.84

%Change (24-0 wk) 22.2 ± 61.7 5.0 ± 45.2 8.6 ± 44.1 0.39 0.18

Insulin Tmax (min) 0.60

Baseline 83.8 ± 34.9 90.9 ± 32.4 85.4 ± 34.7 0.64

Wk 24 85.1 ± 145 92.1 ± 32.7 77.8 ± 35.3 0.12

%Change (24-0 wk) 29.8 ± 137 22.7 ± 84.1 6.4 ± 64.4 0.53 0.36

Area under the curve of 0.39

insulin (pU/L x 120

min x 10-3)c)

Baseline 5241 ± 7829 5639 ± 5386 5335 ± 3652 0.84

Wk 24 5644 ± 6613 5364 ± 4935 5332 ± 3882 0.87

%Change (24-0 wk) 14.3 ± 48.1 3.1 ± 41.6 7.2 ± 42.0 0.57 0.39

HOMA-IR (homeostasis 0.30

model assessment for

insulin resistance)

Baseline 2.1 ± 3.7 2.1 ± 2.5 2.2 ± 1.5 0.90

Wk 24 2.0 ± 1.3 2.1 ± 1.6 1.8 ± 1.5 0.55

%Change (24-0 wk) 22.8 ± 107 21.6 ± 98.9 -0.2 ± 64.0 0.39 0.21

ISlMatsuda (insulin 0.25

sensitivity index)

Baseline 89.7 ± 81.2 85.3 ± 73.7 79.5 ± 56.2 0.67

Wk 24 105.8 ± 83.0 94.9 ± 56.9 107 ± 51.3 0.64

%Change (24-0 wk) 20.5 ± 74.9 9.3 ± 47.5 27.6 ± 56.1 0.39 0.20

HOMA-0 (homeostasis 0.91

model assessment for

B-cell function)

Baseline 73.3 ± 62.3 77.6 ± 84.1 75.7 ± 60.8 0.92

Wk 24 77.9 ± 3.4 84.0 ± 80.2 86.1 ± 109 0.75

%Change (24-0 wk) 43.9 ± 153 31.0 ± 105 25.7 ± 200 0.56 0.83

QUICKI (quantitative 0.21

insulin sensitivity

check index)

Baseline 0.61 ± 0.15 0.6 1± 0.15 0.60 ± 0.11 0.78

Wk 24 0.61 ± 0.13 0.61 ± 0.10 0.63 ± 0.10 0.55

%Change (24-0 wk) 2.32 ± 19.5 0.18 ± 15.6 6.29 ± 17.8 0.34 0.08

a) All of the values presented in this table are unadjusted and were the geometric means ± SD; Cmax, maximum concentration; 7"max, time

to ^max'

b) p1: ANOVA; p2: p-interaction (time x group) by ANOVA for repeated measures; p3: ANCOVA, covariate with adjusted for body weight change at follow-up and baseline (24 wk-baseline).

c) The area under the curve of insulin includes 30, 60, and 120 min postload.

significantly different among the three groups. Small and comparable changes in body weight, body fat, and waist circumference were observed among the treatment groups. The sensitivity analyses also showed similar null effects after further adjusting for the weight changes. Thus, the null effect is unlikely to be due to the imbalance in the baseline values and

the changes in body weight or fat and the above-mentioned lifestyle factors.

Our trial had several features that support the validity of our results. First, the similarity in the baseline characters among the three arms demonstrated successful randomization. Second, we had a very low dropout rate (8.5%) andahigh

rate of supplement consumption (97%). Third, the results of the urinary isoflavones excretion tests suggested the successful uptake of the intervention. All of these factors decrease the possibility of a false null effect due to an inadequate intervention. Moreover, successful blinding helped us to control for potential information bias.

This study has several limitations. As our sample-size calculation was based on the change of the primary outcome of FG, we might not have adequate power to detect significant but small differences in the changes of other outcomes. For example, we only had a power of 54% to detect a mean difference of 11.26% between the Genistein and Placebo groups for FIns. Second, the relatively short duration of the intervention might have limited our ability to detect potential long-term effects.

In conclusion, our findings suggest that daily supplements of 50 mg of isolated daidzein or genistein do not significantly improve glucose control or insulin resistance in patients with IGR over a period of 6 months.

YMC, YBY, and YBL designed the study; YBY, ALC, WL, SYZ, JL, JHG, WPD, and SF conducted the research; and YBY analyzed the data. YBY and YMC wrote the manuscript. YMC and YBY have the primary responsibility for the final content. All ofthe authors have read and approved the final manuscript.

We thank Li-ping He, Juan Deng, Wen-qiong Xue, and Yuan-ting Liu for supplement labeling, and Chun-bo Zhang for isoflavones analysis assistance. This study was jointly supported by the Department of Health of Guangdong Province (no: A2008158); Program for New Century Excellent Talents in University, Ministry of Education, China (NCET-06-0719); Chinese Nutrition Society in 2006 and Danone Institute in 2009.

The authors have declared no conflict ofinterest.

5 References

[1] Whiting, D. R., Guariguata, L., Weil, C., Shaw, J., IDF diabetes atlas: global estimates of the prevalence of diabetes for 2011 and 2030. Diabetes Res. Clin. Pract. 2011, 94, 311-321.

[2] Yang, W., Lu, J., Weng, J., Jia, W. et al., Prevalence of diabetes among men and women in China. N. Engl. J. Med. 2010, 362, 1090-1101.

[3] Alberti, K. G., Zimmet, P, Shaw, J., International Diabetes Federation: a consensus on Type 2 diabetes prevention. Di-abet. Med. 2007, 24, 451-463.

[4] Bartke, A., Peluso, M. R., Moretz, N., Wright, C. et al., Effects of Soy-derived diets on plasma and liver lipids, glucose tolerance, and longevity in normal, long-lived and short-lived mice. Horm. Metab. Res. 2004, 36, 550-558.

[5] Lu, M. P., Wang, R., Song, X., Chibbar, R. et al., Dietary soy isoflavones increase insulin secretion and prevent the development of diabetic cataracts in streptozotocin-induced diabetic rats. Nutr. Res. 2008, 28, 464-471.

[6] Mezei, O., Banz, W. J., Steger, R. W., Peluso, M. R. et al., Soy isoflavones exert antidiabetic and hypolipidemic effects

through the PPAR pathways in obese Zucker rats and murine RAW 264.7 cells. J. Nutr. 2003, 133, 1238-1243.

[7] Yu, Z. W., Eriksson, J. W., The upregulating effect of insulin and vanadate on cell surface insulin receptors in rat adipocytes is modulated by glucose and energy availability. Horm. Metab. Res. 2000, 32, 310-315.

[8] Villegas, R., Gao, Y T., Yang, G., Li, H. L. et al., Legume and soy food intake and the incidence of type 2 diabetes in the Shanghai Women's Health Study. Am. J. Clin. Nutr. 2008, 87, 162-167.

[9] Mueller, N. T., Odegaard, A. O., Gross, M. D., Koh, W. P et al., Soy intake and risk of type 2 diabetes in Chinese Singaporeans [corrected]. Eur. J. Nutr. 2012, 51, 1033-1040.

[10] Morimoto, Y, Steinbrecher, A., Kolonel, L. N., Maskarinec, G., Soy consumption is not protective against diabetes in Hawaii: the Multiethnic Cohort. Eur. J. Clin. Nutr. 2011, 65, 279-282.

[11] Nanri, A., Mizoue, T., Takahashi, Y., Kirii, K. et al., Soy product and isoflavone intakes are associated with a lower risk of type 2 diabetes in overweight Japanese women. J. Nutr. 2010, 140, 580-586.

[12] Ricci, E., Cipriani, S., Chiaffarino, F., Malvezzi, M. et al., Effects of soy isoflavones and genistein on glucose metabolism in perimenopausal and postmenopausal non-Asian women: a meta-analysis of randomized controlled trials. Menopause 2010, 17, 1080-1086.

[13] Liu, Z. M., Chen, Y. M., Ho, S. C., Effects of soy intake on glycemic control: a meta-analysis of randomized controlled trials. Am. J. Clin. Nutr. 2011, 93, 1092-1101.

[14] Villa, P, Costantini, B., Suriano, R., Perri, C. et al., The differential effect of the phytoestrogen genistein on cardiovascular risk factors in postmenopausal women: relationship with the metabolic status. J. Clin. Endocrinol. Metab. 2009, 94, 552-558.

[15] Chen, S. T., Chen, J. R., Yang, C. S., Peng, S. J. et al., Effect of soya protein on serum lipid profile and lipoprotein concentrations in patients undergoing hypercholesterolaemic haemodialysis. Br. J. Nutr. 2006, 95, 366-371.

[16] Wu, A. H., Stanczyk, F. Z., Martinez, C., Tseng, C. C. et al., A controlled 2-mo dietary fat reduction and soy food supplementation study in postmenopausal women. Am. J. Clin. Nutr. 2005, 81, 1133-1141.

[17] Jayagopal, V., Albertazzi, P., Kilpatrick, E. S., Howarth, E. M. et al., Beneficial effects of soy phytoestrogen intake in postmenopausal women with type 2 diabetes. Diabetes Care 2002, 25, 1709-1714.

[18] Liu, Z. M., Chen, Y. M., Ho, S. C., Ho, Y. P et al., Effects of soy protein and isoflavones on glycemic control and insulin sensitivity: a 6-mo double-blind, randomized, placebo-controlled trial in postmenopausal Chinese women with prediabetes or untreated early diabetes. Am. J. Clin. Nutr. 2010, 91, 1394-1401.

[19] Gobert, C. P, Pipe, E. A., Capes, S. E., Darlington, G. A. et al., Soya protein does not affect glycaemic control in adults with type 2 diabetes. Br. J. Nutr. 2010, 103, 412-421.

[20] Okabe, Y., Shimazu, T., Tanimoto, H., Higher bioavailability of isoflavones after a single ingestion of aglycone-rich

fermented soybeans compared with glucoside-rich non-fermented soybeans in Japanese postmenopausal women. J. Sci. FoodAgric. 2011, 91, 658-663.

[21] Izumi, T., Piskula, M. K., Osawa, S., Obata, A. et al., Soy isoflavone aglycones are absorbed faster and in higher amounts than their glucosides in humans. J. Nutr. 2000, 130, 1695-1699.

[22] Weaver, C. M., Cheong, J. M., Soy isoflavones and bone health: the relationship is still unclear. J. Nutr. 2005, 135, 1243-1247.

[23] Morabito, N., Crisafulli, A., Vergara, C., Gaudio, A. et al., Effects of genistein and hormone-replacement therapy on bone loss in early postmenopausal women: a randomized double-blind placebo-controlled study. J. Bone Miner. Res. 2002, 17, 1904-1912.

[24] Marini, H., Minutoli, L., Polito, F, Bitto, A. et al., Effects of the phytoestrogen genistein on bone metabolism in osteopenic postmenopausal women: a randomized trial. Ann. Intern. Med. 2007, 146, 839-847.

[25] Szkudelska, K., Nogowski, L., Genistein — a dietary compound inducing hormonal and metabolic changes. J. Steroid Biochem. Mol. Biol. 2007, 105, 37-45.

[26] Lavigne, C., Marette, A., Jacques, H., Cod and soy proteins compared with casein improve glucose tolerance and insulin sensitivity in rats. Am. J. Physiol. Endocrinol. Metab. 2000, 278, E491-E500.

[27] Choi, M. S., Jung, U. J., Yeo, J., Kim, M. J. et al., Genistein and daidzein prevent diabetes onset by elevating insulin level and altering hepatic gluconeogenic and lipogenic enzyme activities in non-obese diabetic (NOD) mice. Diabetes Metab. Res. Rev. 2008, 24, 74-81.

[28] Crisafulli, A., Altavilla, D., Marini, H., Bitto, A. et al., Effects of the phytoestrogen genistein on cardiovascular risk factors in postmenopausal women. Menopause 2005, 12, 186192.

[29] Atteritano, M., Marini, H., Minutoli, L., Polito, F. et al., Effects of the phytoestrogen genistein on some predictors of cardiovascular risk in osteopenic, postmenopausal women: a two-year randomized, double-blind, placebo-controlled study. J. Clin. Endocrinol. Metab. 2007, 92, 3068-3075.

[30] Marini, H., Bitto, A., Altavilla, D., Burnett, B. P et al., Efficacy of genistein aglycone on some cardiovascular risk factors and homocysteine levels: a follow-up study. Nutr. Metab. Cardiovasc. Dis. 2010, 20, 332-340.

[31] Setchell, K. D. R., Brown, N. M., Lydeking-Olsen, E., The clinical importance of the metabolite equol — a clue to the effectiveness of soy and its isoflavones. J. Nutr. 2002, 132, 3577-3584.

[32] Setchell, K. D., Clerici, C., Lephart, E. D., Cole, S. J. et al., S-equol, a potent ligand for estrogen receptor beta, is the exclusive enantiomeric form of the soy isoflavone metabolite

produced by human intestinal bacterial flora. Am. J. Clin. Nutr. 2005, 81, 1072-1079.

[33] Yang,Y. X. (Ed.), China Food Composition. Peking University Medical Press, Beijing, China 2004.

[34] Guo, K., Zhang, B., Chen, C., Uchiyama, S. et al., Daidzein-metabolising phenotypes in relation to serum lipids and uric acid in adults in Guangzhou, China. Br. J. Nutr. 2010, 104, 118-124.

[35] Ho, S. C., Chen, Y M., Ho, S. S., Woo, J. L., Soy isoflavone supplementation and fasting serum glucose and lipid profile among postmenopausal Chinese women: a double-blind, randomized, placebo-controlled trial. Menopause 2007, 14, 905-912.

[36] Franke, A. A., Lai, J. F, Halm, B. M., Pagano, I. et al., Equol production changes overtime in postmenopausal women. J. Nutr. Biochem. 2012, 23, 573-579.

[37] Mohamad-Shahi, M., Karandish, M., Haidari, F., Omidian, K. et al., Effect of daidzein-low-calorie diet on body weight, serum levels of glucose, resistin, and high sensitive C-reactive protein in high fat, high calorie diet induced rats. Saudi Med. J. 2012, 33, 70-75.

[38] Yang, G., Shu, X. O., Jin, F., Elasy, T. et al., Soyfood consumption and risk of glycosuria: a cross-sectional study within the Shanghai Women's Health Study. Eur. J. Clin. Nutr. 2004, 58, 615-620.

[39] Yang, B., Chen, Y., Xu, T., Yu, Y. et al., Systematic review and meta-analysis of soy products consumption in patients with type 2 diabetes mellitus. Asia Pac. J. Clin. Nutr. 2011, 20, 593-602.

[40] Zhang, Y. B., Chen, W. H., Guo, J. J., Fu, Z. H. et al., Soy isoflavone supplementation could reduce body weight and improve glucose metabolism in non-Asian postmenopausal women — a meta-analysis. Nutrition 2013, 29, 8-14.

[41] Squadrito, F, Marini, H., Bitto, A., Altavilla, D. et al., Genistein in the metabolic syndrome: results of a randomized clinical trial. J. Clin. Endocrinol. Metab. 2013, 98, 3366-3374.

[42] Khaodhiar, L., Ricciotti, H. A., Li, L., Pan, W. et al., Daidzein-rich isoflavone aglycones are potentially effective in reducing hot flashes in menopausal women. Menopause 2008, 15, 125-132.

[43] Jackson, R. L., Greiwe, J. S., Schwen, R. J., Emerging evidence of the health benefits of S-equol, an estrogen receptor beta agonist. Nutr. Rev. 2011, 69, 432-448.

[44] Usui, T., Tochiya, M., Sasaki, Y., Muranaka, K. et al., Effects of natural S-equol supplements on overweight or obesity and metabolic syndrome in the Japanese, based on sex and equol status. Clin. Endocrinol. 2013, 78, 365-372.

[45] Shahbazian, H., Reza, A., Javad, S., Heshmatollah, S. et al., Beneficial effects of soy protein isoflavones on lipid and blood glucose concentrations in type 2 diabetic subjects. Saudi Med. J. 2007, 28, 652-654.