Abstract
Some evidence suggests that phytoestrogens, such as soy-derived isoflavones, may have beneficial effects on cardiovascular health and glycemic control. These data are mainly limited to postmenopausal women or individuals at elevated cardiometabolic risk. There is a lack of data for pregnant women who have elevated estrogen levels and physiologically altered glucose and lipid metabolism. We analyzed data from 299 pregnant women who participated in the NHANES 2001–2008 surveys. Multivariable linear regression analyses were used to examine the association between urinary concentrations of isoflavonoids and cardiometabolic risk markers, adjusted for body mass index, pregnancy trimester, total energy intake, dietary intake of protein, fiber, and cholesterol, and demographic and lifestyle factors. Cardiometabolic risk markers were log-transformed, and geometric means were calculated by quartiles of urinary concentrations of isoflavonoids. Comparing women in the highest vs. lowest quartiles of urine total isoflavone concentrations, we observed significant, inverse associations with circulating concentrations of fasting glucose (79 vs. 88 mg/dL, P-trend = 0.0009), insulin (8.2 vs. 12.8 μU/mL, P-trend = 0.03), and triglyceride (156 vs. 185 mg/dL, P-trend = 0.02), and the homeostasis model assessment of insulin resistance (1.6 vs. 2.8, P-trend = 0.01), but not for total, low-density lipoprotein cholesterol and high-density lipoprotein cholesterol. The concentrations of individual isoflavonoids, daidzein, equol, and O-desmethylangolensin were inversely associated with some cardiometabolic risk markers, although no clear pattern emerged. These data suggest that there may be a relation between isoflavone intake and cardiometabolic risk markers in pregnant women.
Introduction
Soy isoflavones, a plant-based class of phytoestrogens, have received considerable attention for their potential anticarcinogenic, cardioprotective, and antioxidant properties (1). Isoflavones consist primarily of the glycosides daidzin and genistin, which are hydrolyzed into the bioactive aglycones daidzein and genistein through bacterial metabolism in the small intestine (2, 3). Intestinal microbiota can further metabolize daidzein into the glycoside conjugates of equol or O-desmethylangolensin (O-DMA)8. Either the intact or modified phytoestrogens are absorbed by diffusion into circulation (1). There are conflicting data with regard to the role of isoflavones in lipid and glucose metabolic regulation, which highlights the importance of further characterizing potential relations. Were a relation to exist, a potential mechanism involves the estrogen-mimicking characteristics of isoflavone. The chemical properties of isoflavones are similar to estradiol, both agonizing or antagonizing the α and β estrogen receptors (4, 5).
There are some data suggesting that soy isoflavones have beneficial effects on lipid homeostasis and glycemic control (6–10). However, these data are mainly limited to postmenopausal women or individuals at elevated cardiometabolic risk, such as patients with type 2 diabetes. Little is known regarding the effects of isoflavone intake on cardiometabolic health in pregnant women; women at elevated risk of developing glucose intolerance and dyslipidemia, partially attributable to elevated estrogen concentrations during pregnancy, may thus particularly benefit from isoflavone intake. In addition, isoflavones can diffuse across the placenta and reach the fetus (11, 12). It has been postulated that soy isoflavones in fetal circulation may reduce the susceptibility to cardiometabolic disorders in adulthood (11, 13). Animal studies have revealed the beneficial effects of soy isoflavones on fetal programming. For example, the offspring of mice fed isoflavones during pregnancy had a reduced risk of obesity and metabolic disorders in later life (14, 15). Therefore, it is important to assess the relation between soy isoflavones and markers of lipid and glucose metabolism in pregnant women.
In the current study, we use data from the NHANES to examine potential associations between urinary isoflavone metabolite concentrations, biomarkers of dietary isoflavone intake, and cardiometabolic risk markers in pregnant women.
Methods
Study population.
The NHANES is a cross-sectional study among the U.S. civilian population and is designed to assess health and nutritional status through interviews, along with physical and laboratory examinations (16). A complex, multistaged, stratified, clustered sampling method is used in the NHANES, with oversampling of adolescents aged 12–19 y, older Americans aged ≥60 y, African Americans, and Mexican Americans. In the current study, we used data from 4 NHANES, 2001–2002, 2003–2004, 2005–2006, and 2007–2008, in which urinary concentrations of isoflavones were measured. Each survey enrolled ∼10,000 individuals evenly distributed between women and men (11,039 individuals in the 2001–2002 survey, 10,122 in the 2003–2004 survey, 10,348 in the 2005–2006 survey, and 10,149 in the 2007–2008 survey) (17–20). All participants provided written informed consent, and the study protocol was approved by the institutional review board at the CDC.
In the current study, we restricted our analysis to pregnant women who completed the FFQ and had valid urinary isoflavone concentrations (n = 299). The sample sizes varied for cardiometabolic risk biomarkers: n = 278 for total cholesterol (TC), 116 for LDL-cholesterol (LDL-C), 285 for HDL-cholesterol (HDL-C), 126 for TGs and the TG:HDL-C ratio, 128 for glucose, 127 for insulin, and 127 for HOMA-IR. We included individuals who had non-missing data for each of these markers to preserve statistical power.
Measurements.
Blood was collected by venipuncture, and spot urine samples were collected in the NHANES mobile examination centers. Detailed specimen collection, processing, and testing information is available on the NHANES Web site (16). Urinary concentrations of isoflavonoids, including genistein, daidzein, equol, and O-DMA, were measured by HPLC–electrospray ionization–MS/MS (21). Total isoflavones included daidzein, genistein, O-DMA, and equol. In data analyses, the concentrations of isoflavonoids were adjusted for urinary creatinine to account for variability in urine dilution. Cardiometabolic risk markers were measured as described previously (16). Insulin resistance was estimated using HOMA-IR, which is equal to (fasting insulin in μU/mL × fasting glucose in mmol/L)/22.5 (22).
Pregnancy tests were performed on female participants aged 12–59 y and menstruating females aged 8–11 y by a human chorionic gonadotropin test kit (Icon 25 hCG test kit; Beckman Coulter) to detect change in human chorionic gonadotropin in urine or serum. Physical examination, including weight, height, waist circumference, and blood pressure measurements, were performed at mobile examination centers following standard protocols. BMI was calculated as weight in kilograms divided by height in meters squared. BMI was based on the data that was available, and pregnant weight was measured at the time of the survey.
Dietary intake of soy foods was assessed by frequency of consuming soy milk and soy foods, such as tofu, cooked dried soy beans, soy burgers, or soy meat substitutes, using a self-administered FFQ designed to capture data during the past year (23). For each food item, participants were asked how often they consumed each food item with prespecified portion sizes. The Diet*Calc software was used to produce daily food frequency estimates from the FFQ data. Because FFQ data does not capture servings for each individual soy-containing food, we were unable to obtain the amount of dietary intake of isoflavones from the FFQ. The questions in the FFQ remained the same across all of the years of data collection. For total energy intake and dietary intake of nutrients, the average of intakes from 2-d, 24-h dietary recall was used in the 2003–2008 cycles, and data from a single 24-h recall was used in the 2001–2002 data cycle.
Age, race/ethnicity, education, pregnancy trimester, alcohol use, cigarette smoking, and physical activity were self-reported. Race/ethnicity was categorized as non-Hispanic white, non-Hispanic black, Mexican American, other Hispanic, and other race/ethnicity. Educational attainment was categorized as less than high school graduate, high school graduate, and college or higher (which includes some college with no degree, associate’s degree, bachelor’s degree, and above). Smoking status was determined by the question on frequency of current cigarette smoking. Women who smoked every day or some days at the time of the survey were coded as being current smokers. Alcohol use was determined by the question on the average number of alcoholic drinks per day in the past 12 mo. Moderate-to-vigorous physical activity (MVPA) was defined as having moderate or vigorous physical activity over the past 30 d.
Statistical analyses.
The complex sampling design of the NHANES was taken into account in statistical analyses. In descriptive analyses, arithmetic mean and weighted SE were reported for continuous variables, and counts and weighted percentages were reported for categorical variables. Weighted median and 95% CI were used to describe concentrations of urinary isoflavonoids and cardiometabolic risk markers that had skewed distributions.
χ2 test and ANOVA were used when appropriate to examine differences in characteristics of participants across quartiles of urinary isoflavone metabolite concentrations. Multivariable linear regression analyses were used to evaluate the association between cardiometabolic risk markers and creatinine-adjusted urinary concentrations of isoflavones. Analyses were adjusted for demographic characteristics (age, education, race/ethnicity), BMI, pregnancy trimester, total energy intake, nutrient intake, and lifestyle factors (alcohol consumption, cigarette smoking, and MVPA). Cardiometabolic risk markers were log-transformed to normalize the distribution of these variables. Geometric means and 95% CIs were obtained based on the least-square means estimated by quartiles of urinary isoflavone metabolite concentrations. The P-trend was tested by modeling the median value for each quartile of urinary isoflavone metabolites as a continuous variable. To examine the correlation between dietary soy intake and urinary isoflavonoids, sample-weighted partial Pearson’s correlation coefficients were calculated. In addition, multivariable-adjusted means of urinary excretion of isoflavonoids were calculated by frequency of soy food intake, and P-trend was tested. All analyses were performed using SAS software (version 9.2; SAS Institute). Statistical significance for all tests was established at α = 0.05.
Results
The mean age of the participants was 28 y (Table 1). The majority were non-Hispanic white, followed by Mexican American, non-Hispanic black, other ethnicity, and other Hispanic ethnicity. Most of the participants attained a college degree or higher. A small proportion (5.4%) of the pregnant women was current smokers. Sixty-two percent of the pregnant women had MVPA, and nearly half of the sample did not drink alcohol in the past year. The participants’ demographic characteristics varied by urinary concentrations of isoflavones. Compared with the participants in the lowest quartile of urinary total isoflavones, those in the highest quartile of total isoflavones were older (P = 0.01), had higher educational attainment (P = 0.01), and were more likely to be non-Hispanic white (P = 0.001). Participants in the highest quartile of total urinary isoflavone excretion also consumed more protein (P = 0.01), fiber (P = 0.008), and cholesterol (P = 0.02) compared with those in the lowest quartile (Table 2).
TABLE 1.
| Quartile of total isoflavones |
|||||
| Characteristics | 1 (n = 74) | 2 (n = 75) | 3 (n = 75) | 4 (n = 75) | Total (n = 299) |
| Age, y | 27 ± 0.8 | 26 ± 0.6 | 29 ± 0.6 | 30 ± 0.5 | 28 ± 0.4 |
| BMI, kg/m2 | 30.1 ± 1.1 | 27.3 ± 0.6 | 28.3 ± 0.4 | 26.7 ± 0.8 | 28.2 ± 0.5 |
| Race/ethnicity | |||||
| Non-Hispanic white | 25 (49.5) | 28 (37.8) | 41 (67.5) | 42 (68.8) | 136 (55.3) |
| Non-Hispanic black | 15 (20.0) | 13 (20.1) | 6 (6.2) | 5 (4.9) | 39 (13.2) |
| Mexican American | 29 (28.2) | 22 (22.2) | 24 (15.6) | 16 (16.7) | 91 (21.0) |
| Other Hispanic ethnicity | 3 (1.3) | 7 (8.8) | 1 (2.2) | 2 (3.2) | 13 (3.8) |
| Other race/ethnicity | 2 (1.1) | 5 (11.1) | 3 (8.4) | 10 (6.5) | 20 (6.6) |
| Education | |||||
| Less than high school graduate | 21 (24.0) | 20 (21.0) | 21 (20.5) | 14 (11.6) | 76 (19.7) |
| High school graduate | 25 (25.8) | 22 (23.6) | 14 (14.7) | 11 (11.1) | 72 (19.3) |
| College and above3 | 28 (50.1) | 33 (55.4) | 40 (64.8) | 50 (77.3) | 151 (61.0) |
| Pregnancy trimester | |||||
| 1st | 17 (26.1) | 10 (16.9) | 3 (7.0) | 16 (27.8) | 46 (19.3) |
| 2nd | 20 (21.2) | 25 (30.0) | 31 (33.7) | 29 (45.2) | 105 (31.7) |
| 3rd | 19 (21.6) | 29 (34.9) | 31 (29.0) | 24 (23.1) | 103 (27.1) |
| Unknown | 18 (31.1) | 11 (18.3) | 10 (30.3) | 6 (3.9) | 45 (21.9) |
| Current smoker | 4 (7.7) | 7 (5.9) | 6 (3.3) | 5 (4.3) | 22 (5.4) |
| Alcohol use in the past year4 | |||||
| None | 41 (45.9) | 44 (53.7) | 34 (49.0) | 35 (33.1) | 154 (45.9) |
| 1–2 drinks/d | 20 (32.3) | 20 (33.1) | 25 (29.8) | 33 (59.3) | 98 (37.6) |
| ≥3 drinks/d | 13 (21.9) | 11 (13.3) | 16 (21.1) | 7 (7.6) | 47 (16.5) |
| MVPA | 38 (59.1) | 30 (47.0) | 42 (66.8) | 54 (77.6) | 164 (62.0) |
Unweighted number of participants. MVPA, moderate-to-vigorous physical activity.
Values are means ± weighted SEs for continuous variables and n (%) for categorical variables, for which n was the actual number of participants included in the analysis and percentage was weighted estimates.
Includes some college with no degree, associate’s degree, bachelor’s degree, and above.
One drink is equivalent to a 12-oz beer, a 5-oz glass of wine, or 1.5 oz of liquor. It contains about 0.6 fluid oz or 14 g of pure ethanol alcohol.
TABLE 2.
| Quartile of total isoflavones |
|||||
| Dietary intake | 1 (n = 69) | 2 (n = 72) | 3 (n = 74) | 4 (n = 71) | Total (n = 286) |
| Energy, Mcal/d | 2.29 ± 0.10 | 2.36 ± 0.06 | 2.23 ± 0.03 | 2.56 ± 0.09 | 2.35 ± 0.06 |
| Carbohydrate, g/d | 320 ± 15 | 313 ± 8 | 296 ± 8 | 340 ± 15 | 317 ± 9 |
| Protein, g/d | 78 ± 5 | 87 ± 4 | 88 ± 5 | 98 ± 4 | 87 ± 3* |
| Dietary fiber, g/d | 17 ± 1 | 19 ± 1 | 18 ± 1 | 23 ± 1 | 19 ± 1* |
| Total sugars, g/d | 155 ± 8 | 152 ± 5 | 143 ± 7 | 156 ± 9 | 151 ± 6 |
| Total fat, g/d | 80 ± 4 | 88 ± 3 | 81 ± 3 | 94 ± 3 | 86 ± 3 |
| SFAs, g/d | 28 ± 2 | 30 ± 1 | 26 ± 1 | 33 ± 1 | 29 ± 1 |
| MUFAs, g/d | 30 ± 2 | 33 ± 1 | 30 ± 1 | 35 ± 1 | 32 ± 1 |
| PUFAs, g/d | 17 ± 1 | 18 ± 1 | 18 ± 1 | 19 ± 1 | 18 ± 1 |
| Cholesterol, mg/d | 253 ± 39 | 281 ± 22 | 323 ± 21 | 311 ± 18 | 291 ± 14* |
| Sugar-sweetened beverage, g/d | 696 ± 88 | 569 ± 56 | 459 ± 37 | 654 ± 24 | 595 ± 35 |
| Fruit, g/d | 189 ± 13 | 254 ± 7 | 213 ± 20 | 176 ± 13 | 209 ± 10 |
| Vegetables, g/d | 111 ± 8 | 176 ± 11 | 257 ± 25 | 145 ± 17 | 174 ± 14 |
Unweighted number of participants.
Values are means ± weighted SEs. *P < 0.05 (ANOVA).
The weighted median urinary total isoflavonoid concentration was 502 (95% CI: 260, 745) μg/g creatinine, and those of individual isoflavonoids were 48 (95% CI: 24, 73) μg daidzein/g creatinine, 22 (95% CI: 15, 29) μg genistein/g creatinine, 4 (95% CI: 3, 5) μg O-DMA/g creatinine, and 8 (95% CI: 6, 10) μg equol/g creatinine. The weighted median values of cardiometabolic risk markers were 211 (95% CI: 202, 219) mg/dL for TC, 110 (95% CI: 101, 118) mg/dL for LDL-C, 60 (95% CI: 58, 62) mg/dL for HDL-C, 159 (95% CI: 147, 170) mg/dL for TG, 2.7 (95% CI: 2.5, 3.0) mg/dL for TG:HDL-C ratio, 85 (95% CI: 83, 87) mg/dL for fasting glucose, 15.0 (95% CI: 12.2, 17.8) μU/mL for fasting insulin, and 3.4 (95% CI: 2.7, 4.0) for HOMA-IR.
Self-reported soy intake frequency was positively correlated with urinary isoflavone and its metabolite concentrations (Pearson's partial correlation, r = 0.4, P < 0.001 for total isoflavones; r = 0.3, P < 0.001 for daidzein; r = 0.3, P < 0.001 for O-DMA; r = 0.2, P < 0.001 for equol; and r = 0.3, P < 0.001 for genistein). Significant increasing trends were observed between urinary excretion of isoflavonoids and soy food intake frequency (Fig. 1; P-trend = 0.0009 for daidzein, P-trend = 0.006 for equol, P-trend < 0.0001 for O-DMA, and P-trend = 0.0001 for genistein, respectively).
FIGURE 1.
Urinary excretion of daidzein (A); equol (B); O-desmethylangolensin (C); and genistein (D) by levels of soy food intake among pregnant women, NHANES 2001–2008. Values are multivariable-adjusted means and 95% CIs; n = 125•••. Data were adjusted for age at urine sample collection (y), BMI (kg/m2), total energy intake (Kcal), dietary intake of protein (g), dietary fiber (g), dietary intake of cholesterol (g), current smoker (yes, no), alcohol use (none, 1–2 drinks/d, ≥3 drinks/d), engaging in moderate-to-vigorous physical activity (yes, no), ethnicity (non-Hispanic white, non-Hispanic black, Mexican American, other Hispanic groups, and other ethnicity), month of pregnancy (1–3 trimesters or unknown).
Table 3 presents the multivariable-adjusted geometric means for cardiometabolic risk markers by quartiles of urinary concentrations of isoflavonoids. In general, there was a significant inverse relation between concentrations of urinary isoflavonoids and some cardiometabolic risk markers: comparing women in the highest vs. lowest quartiles of total isoflavone concentration, multivariable-adjusted fasting glucose concentrations were 79 vs. 88 mg/dL (P-trend = 0.0009), 8.2 vs. 12.8 μU/mL (P-trend = 0.03) for fasting insulin concentrations, 1.6 vs. 2.8 (P-trend = 0.01) for HOMA-IR, and 156 vs. 185 mg/dL (P-trend = 0.02) for TG.
TABLE 3.
Multivariable-adjusted means of metabolic risk markers by urinary concentrations of isoflavone metabolites among pregnant women, NHANES 2001–20081
| Quartiles of urinary markers |
|||||
| 1 | 2 | 3 | 4 | P-trend | |
| Total isoflavones | |||||
| TC, mg/dL | 224 (213, 235) | 209 (201, 218) | 199 (189, 209) | 215 (203, 228) | 0.13 |
| LDL-C, mg/dL | 104 (91, 118) | 109 (99, 120) | 96 (85, 108) | 124 (110, 138) | 0.16 |
| HDL-C, mg/dL | 61 (56, 67) | 59 (54, 65) | 55 (52, 59) | 61 (57, 65) | 0.62 |
| TG, mg/dL | 185 (164, 209) | 160 (139, 183) | 135 (114, 159) | 156 (141, 174) | 0.02 |
| TG:HDL-C ratio | 2.9 (2.5, 3.4) | 2.7 (2.3, 3.3) | 2.4 (2.0, 2.9) | 2.5 (2.2, 2.9) | 0.12 |
| Fasting glucose, mg/dL | 88 (85, 91) | 86 (83, 89) | 84 (81, 87) | 79 (76, 83) | 0.0009 |
| Fasting insulin, μU/mL | 12.8 (9.8, 16.6) | 11.6 (9.0, 15.0) | 11.1 (8.2, 15.1) | 8.2 (6.4, 10.5) | 0.03 |
| HOMA-IR | 2.8 (2.1, 3.7) | 2.5 (1.9, 3.3) | 2.3 (1.7, 3.2) | 1.6 (1.2, 2.1) | 0.01 |
| Daidzein | |||||
| TC, mg/dL | 230 (218, 242) | 208 (202, 215) | 206 (197, 215) | 202 (188, 217) | 0.01 |
| LDL-C, mg/dL | 114 (98, 133) | 102 (95, 111) | 106 (95, 118) | 114 (98, 132) | 0.80 |
| HDL-C, mg/dL | 63 (58, 68) | 56 (52, 61) | 60 (55, 65) | 58 (54, 62) | 0.46 |
| TG, mg/dL | 169 (139, 205) | 158 (140, 179) | 166 (146, 190) | 140 (120, 164) | 0.22 |
| TG:HDL-C ratio | 2.7 (2.2, 3.3) | 2.7 (2.0, 3.2) | 2.8 (2.4, 3.3) | 2.3 (2.0, 2.7) | 0.39 |
| Fasting glucose, mg/dL | 83 (80, 87) | 88 (85, 91) | 82 (79, 85) | 84 (81, 87) | 0.37 |
| Fasting insulin, μU/mL | 9.5 (7.5, 12.1) | 13.5 (10.7, 17.1) | 11.3 (8.7, 14.6) | 8.5 (6.8, 10.7) | 0.20 |
| HOMA-IR | 2.0 (1.5, 2.5) | 2.9 (2.3, 3.8) | 2.3 (1.7, 3.0) | 1.8 (1.4, 2.3) | 0.18 |
| Equol | |||||
| TC, mg/dL | 217 (208, 227) | 221 (210, 232) | 205 (195, 216) | 202 (191, 214) | 0.04 |
| LDL-C, mg/dL | 113 (101, 126) | 118 (103, 134) | 107 (94, 123) | 101 (90, 114) | 0.14 |
| HDL-C, mg/dL | 60 (55, 65) | 57 (54, 60) | 63 (57, 69) | 57 (53, 61) | 0.47 |
| TG, mg/dL | 194 (170, 221) | 155 (130, 185) | 162 (142, 1851) | 143 (125, 165) | 0.004 |
| TG:HDL-C ratio | 3.0 (2.5, 3.6) | 2.6 (2.2, 3.2) | 2.7 (2.2, 3.2) | 2.5 (2.1, 3.0) | 0.18 |
| Fasting glucose, mg/dL | 87 (84, 90) | 83 (79, 88) | 85 (83, 87) | 83 (79, 86) | 0.04 |
| Fasting insulin, μU/mL | 9.6 (7.4, 12.4) | 9.4 (6.8, 13.0) | 11.3 (9.0, 14.2) | 11.8 (9.0, 15.5) | 0.31 |
| HOMA-IR | 2.1 (1.6, 2.7) | 1.9 (1.3, 2.8) | 2.4 (1.9, 3.0) | 2.4 (1.8, 3.2) | 0.47 |
| O-DMA | |||||
| TC, mg/dL | 222 (211, 233) | 211 (203, 220) | 211 (200, 223) | 201 (191, 212) | 0.01 |
| LDL-C, mg/dL | 121 (109, 134) | 105 (92, 120) | 96 (86, 108) | 111 (101, 122) | 0.38 |
| HDL-C, mg/dL | 59 (56, 62) | 59 (54, 64) | 59 (54, 64) | 60 (55, 65) | 0.74 |
| TG, mg/dL | 164 (140, 193) | 173 (147, 204) | 166 (148, 186) | 139 (123, 157) | 0.04 |
| TG:HDL-C ratio | 2.6 (2.2, 3.2) | 3.1 (2.6, 3.7) | 2.8 (2.5, 3.1) | 2.3 (2.0, 2.6) | 0.11 |
| Fasting glucose, mg/dL | 82 (77, 86) | 87 (84, 91) | 86 (83, 89) | 82 (80, 84) | 0.91 |
| Fasting insulin, μU/mL | 9.3 (7.5, 11.5) | 12.8 (9.3, 17.4) | 14.4 (11.3, 18.3) | 8.0 (6.5, 9.9) | 0.22 |
| HOMA-IR | 1.9 (1.5, 2.4) | 2.8 (1.9, 3.9) | 3.1 (2.4, 4.0) | 1.6 (1.3, 2.0) | 0.25 |
| Genistein | |||||
| TC, mg/dL | 218 (207, 228) | 217 (208, 226) | 195 (186, 204) | 216 (201, 233) | 0.45 |
| LDL-C, mg/dL | 110 (95, 128) | 106 (96, 117) | 97 (86, 109) | 124 (110, 139) | 0.18 |
| HDL-C, mg/dL | 62 (57, 67) | 57 (53, 60) | 59 (54, 64) | 59 (55, 64) | 0.68 |
| TG, mg/dL | 171 (147, 200) | 155 (135, 178) | 145 (127, 167) | 171 (155, 188) | 0.98 |
| TG:HDL-C ratio | 2.7 (2.2, 3.2) | 2.6 (2.2, 3.0) | 2.6 (2.2, 3.0) | 2.8 (2.5, 3.2) | 0.55 |
| Fasting glucose, mg/dL | 85 (82, 89) | 86 (83, 89) | 83 (80, 86) | 81 (77, 86) | 0.07 |
| Fasting insulin, μU/mL | 10.8 (7.9, 14.6) | 11.8 (9.0, 15.4) | 10.3 (7.5, 14.3) | 9.6 (7.8, 11.8) | 0.36 |
| HOMA-IR | 2.3 (1.6, 3.2) | 2.5 (1.9, 3.4) | 2.1 (1.5, 3.0) | 1.9 (1.5, 2.4) | 0.27 |
Values are geometric means (95% CIs). Values were adjusted for age at urine sample collection (y), body mass index (kg/m2), total energy intake (Kcal), dietary intake of protein (g), dietary fiber (g), dietary intake of cholesterol (g), current smoker (yes, no), alcohol use (none, 1–2 drinks/d, ≥3 drinks/d), engaging in moderate-to-vigorous physical activity (yes, no), ethnicity (non-Hispanic white, non-Hispanic black, Mexican American, other Hispanic group, and other ethnicity), month of pregnancy (1–3 trimester or unknown). HDL-C, HDL cholesterol; LDL-C, LDL cholesterol; O-DMA, O-desmethylangolensin; TC, total cholesterol.
The concentrations of individual isoflavonoids were also significantly inversely associated with some cardiometabolic risk markers, although no clear patterns emerged. Equol was associated with lower TG and TC when comparing the highest quartile with the lowest [143 vs. 194 mg/dL (P-trend = 0.004) for TG and 202 vs. 217 mg/dL (P-trend = 0.04) for TC]. A significant inverse association was also observed between equol and fasting glucose when comparing the highest quartile with the lowest [83 vs. 87 mg/dL (P-trend = 0.04)]. Daidzein was only associated with lower TC concentrations in the highest quartile (202 mg/dL) compared with the lowest quartile (230 mg/dL) (P-trend = 0.01). O-DMA concentrations were associated with lower TG (P-trend = 0.04) and TC (P-trend = 0.01) concentrations. Genestein was not associated with any cardiometabolic risk markers.
Discussion
In this population-based, cross-sectional investigation using NHANES data, we found that urinary concentrations of isoflavones were associated with a favorable profile of some lipids and markers of insulin sensitivity in pregnant women. To our knowledge, this is the first population-based study examining the association between urinary isoflavones and cardiometabolic risk markers explicitly among pregnant women.
Previous studies, primarily among postmenopausal women and patients with elevated cardiometabolic risks, have provided some evidence supporting the beneficial effect of soy intake on lipid and glucose metabolism. Several meta-analyses on the relation between soy consumption and serum lipids revealed that dietary soy intake may improve lipid profiles by modestly decreasing TC, TG, and LDL-C and increasing HDL-C concentrations (24–29). A recently published meta-analysis on soy isoflavones and glucose metabolism evaluated 10 randomized controlled trials conducted among perimenopausal and postmenopausal non-Asian women and found that soy isoflavones have a beneficial effect on glucose metabolism (30). Although evidence on the benefits of isoflavones on cardiometabolic risk markers appears to be mounting, research specifically targeting pregnant women currently does not exist in the literature. It is known that plasma TG, cholesterol, and HDL-C concentrations increase progressively in normal pregnancy (31–33) and that insulin secretion has to be increased by 2–4 times during the 3rd trimester to compensate for the increased resistance to insulin attributable to hormonal changes during pregnancy (34). Therefore, pregnant women are at high risk of glucose intolerance and hyperlipidemia. Gestational diabetes mellitus (GDM) is the most common medical complication of pregnancy; ∼200,000 or 7% of pregnant women in the United States develop GDM every year (35). Overweight and obese women are at substantially higher risk of developing GDM than women of normal weight (36). Given altered pregnancy physiology and adaptive metabolic derangements as a result of hormonal changes that preferentially support fetal needs, investigation of the associations between isoflavone and cardiometabolic risk markers during pregnancy is warranted. This study found a significant negative relation between urinary isoflavones and some cardiometabolic risk markers in pregnant women. Although the patterns and strengths of associations between individual urinary isoflavonoids and cardiometabolic risk markers varied, this study provided preliminary data on which additional research can be justified and based.
Several potential mechanisms have been proposed by which soy isoflavones may improve lipid and glucose metabolism. As phytoestrogens, isoflavones can compete with endogenous estrogen in binding to estrogen receptors α and β and thus modulate the effects of estrogens on glucose and adipose metabolism (37). It has also been suggested that isoflavones may directly affect pancreatic β-cell function. In vitro and in vivo studies show that genistein stimulates insulin secretion in the insulin-secreting cell lines INS-1 and MIN6 (38, 39) and in cultured islets of Langerhans from mice (38, 40). Genistein and daidzein increase the insulin:glucagon ratio and C-peptide concentration while preserving pancreatic β-cells in non-obese, diabetic mice (41). In addition, isoflavones can bind to PPAR-α and PPAR-γ, which are nuclear receptors that participate in cellular lipid homeostasis and insulin action (42). Soy isoflavones have been found to inhibit lipogenesis and increase lipolysis in the liver and adipocytes, thereby reducing adiposity and improving insulin resistance (43). In this study, we observed beneficial associations between soy isoflavones and some cardiometabolic risk markers in pregnant women. Because hormonal changes during pregnancy are known to alter glucose and lipid metabolism in an unfavorable manner, it is therefore desirable to explore how isoflavones as phytoestrogens can exert effects on lipid and glucose metabolism during pregnancy. It is also important to determine whether the effects vary among populations with different intake levels of soy.
The urinary concentrations of isoflavonoids among pregnant women in this study were comparable with those of the general U.S. population. The average urinary concentration of daidzein among pregnant women in the current study was 0.25 nmol/mg creatinine, 0.11 nmol/mg for genistein, 0.03 nmol/mg for equol, and 0.01 nmol/mg for O-DMA. These figures were close to those reported in a study using NHANES 2001–2004 surveys, which found that the average concentration of urinary daidzein in U.S. adults was 0.23 nmol/mg creatinine, 0.11 nmol/mg for genistein, 0.05 nmol/mg for equol, and 0.03 nmol/mg for O-DMA (44). It is worth noticing that soy intake levels in U.S. populations are much lower than in Asian populations. For example, a study in Shanghai found that the median urinary concentration of daidzein in middle-aged and elderly Chinese women was 5.57 nmol/mg creatinine, 2.41 nmol/mg for genistein, 0.11 nmol/mg for equol, and 1.52 nmol/mg for O-DMA (45). Therefore, it is particularly intriguing to find that soy intake, even at relatively low levels, may still confer beneficial effects.
Equol, an intestinal bacterial metabolite of daidzein, is an important bioactive form of isoflavone. It appears to be superior to the parent isoflavones in binding affinity to estrogen receptors and antioxidant potencies (46). Only approximately one-third of the population of Western countries excrete equol in urine after consuming soy foods (known as “equol producers”), significantly lower than the reported one-half of equol-producers in Asian populations (46–49). In this study, only 22% of participants were classified as equol producers (i.e., urinary equol concentration > 1000 nmol/L) (50). It was found that equol was associated with lower TG, TC, and fasting glucose. Compared with other isoflavonoids, equol had the strongest association with lipid and glucose metabolism.
This study has several strengths and limitations. Regarding strengths, we used data from the NHANES, a large, population-based study with study protocols and procedures that are carefully designed and well-implemented. Soy isoflavone intake was measured by urinary excretion isoflavonoids, which objectively reflects biologically effective doses of isoflavones. It is challenging to capture isoflavone intake from all food and drink sources through an FFQ or 24-h recall because of the ubiquity of soy products in diets (51). In addition, significant variation exists between an individual’s ability to produce equol and O-DMA (52, 53). Using urinary isoflavonoids can capture bioavailable levels of equol and O-DMA in human bodies. Finally, statistical analyses were carefully performed in this study to account for the complex sampling design and to control for potential confounders when investigating the association between isoflavone concentrations and cardiometabolic risk markers.
Regarding limitations, this study was cross-sectional, precluding the possibility of establishing a causal relation between isoflavone concentration and cardiometabolic risks. The small sample size limits the ability to consider ethnicity-specific or trimester-specific analyses. Combining data from multiple cycles introduces the possibility of batch variability, although such variability usually attenuates true associations. In addition, we cannot exclude the possibility of residual and unmeasured confounding. For example, urinary isoflavones were higher among older women, better educated women, and non-Hispanic white women. This finding is consistent with previous studies (45, 54). It was observed in several studies that urinary isoflavonoids were higher among women who engaged in regular exercise and were non-smokers (45, 54, 55). We found similar patterns in this study: those in the highest quartile of total urinary isoflavones had lower BMI, were less likely to be a current smoker, and reported more physical activity, although the differences were not statistically significant. Similarly, TG may have been influenced by MVPA, BMI, and alcohol consumption. Although we controlled for these factors, we cannot rule out the possibility that there was some residual or unmeasured confounding that contributed to the associations observed. Moreover, urinary isoflavone data were only available from a single spot urine sample, reflecting recent intake attributable to its short half-life (3–10 h) (56). However, it was reported that urinary isoflavone concentrations could remain relatively stable, and isoflavone concentrations from a single spot urine sample could reflect the concentration of isoflavones in the body over a 1-y period (57). In this study, we examined the correlation between urinary isoflavone concentration, which measures recent soy intake, and diet intake from FFQ, which measures dietary consumption in the past year. Significant correlations were found between urinary excretion concentration, the short-term intake marker, and soy intake from the FFQ, the relatively long-term measurement of diet intake, suggesting that at least in this population soy intake levels are probably stable over an extended period of time. The results are also in agreement with previous studies that have shown that urinary isoflavones are significantly associated with long-term dietary isoflavone intake (44, 58).
In conclusion, the results of our study showed that urinary excretion of isoflavones was associated with lower circulating concentrations of some cardiometabolic risk markers in pregnant women, a subgroup of the population at increased risk of glucose intolerance and hyperlipidemia. Results from this study may validate associations in pregnancy currently confirmed in other populations. Our findings of varying degrees of associations between individual isoflavonoids and cardiometabolic risk markers supports additional investigation on the soy–cardiometabolic risk association among pregnant women.
Acknowledgments
L.S., Q.S., and L.L.H. designed the research; L.S. and Q.S. analyzed the data; L.S. and H.H.R. conducted the literature search; L.S., H.H.R., E.J., T.A.M.S., A.H.L., Q.S., and L.L.H. wrote the paper; and L.S. had primary responsibility for the final content. All authors read and approved the final manuscript.
Footnotes
Abbreviations used: GDM, gestational diabetes mellitus; HDL-C, HDL-cholesterol; LDL-C, LDL-cholesterol; MVPA, moderate-to-vigorous physical activity; O-DMA, O-desmethylangolensin; TC, total cholesterol.
Literature Cited
- 1.Cederroth CR, Zimmermann C, Nef S. Soy, phytoestrogens and their impact on reproductive health. Mol Cell Endocrinol. 2012;355:192–200. [DOI] [PubMed] [Google Scholar]
- 2.Cederroth CR, Nef S. Soy, phytoestrogens and metabolism: a review. Mol Cell Endocrinol. 2009;304:30–42. [DOI] [PubMed] [Google Scholar]
- 3.Setchell KD. Phytoestrogens: the biochemistry, physiology, and implications for human health of soy isoflavones. Am J Clin Nutr. 1998;68:1333S–46S. [DOI] [PubMed] [Google Scholar]
- 4.Miksicek RJ. Commonly occurring plant flavonoids have estrogenic activity. Mol Pharmacol. 1993;44:37–43. [PubMed] [Google Scholar]
- 5.Kuiper GG, Lemmen JG, Carlsson B, Corton JC, Safe SH, Van Der Saag PT, Van Der Burg B, Gustafsson JA. Interaction of estrogenic chemicals and phytoestrogens with estrogen receptor beta. Endocrinology. 1998;139:4252–63. [DOI] [PubMed] [Google Scholar]
- 6.Yang G, Shu XO, Jin F, Elasy T, Li HL, Li Q, Huang F, Zhang XL, Gao YT, Zheng W. Soyfood consumption and risk of glycosuria: a cross-sectional study within the Shanghai Women's Health Study. Eur J Clin Nutr. 2004;58:615–20. [DOI] [PubMed] [Google Scholar]
- 7.Goodman-Gruen D, Kritz-Silverstein D. Usual dietary isoflavone intake is associated with cardiovascular disease risk factors in postmenopausal women. J Nutr. 2001;131:1202–6. [DOI] [PubMed] [Google Scholar]
- 8.Nanri A, Mizoue T, Takahashi Y, Kirii K, Inoue M, Noda M, Tsugane S. Soy product and isoflavone intakes are associated with a lower risk of type 2 diabetes in overweight Japanese women. J Nutr. 2010;140:580–6. [DOI] [PubMed] [Google Scholar]
- 9.Jayagopal V, Albertazzi P, Kilpatrick ES, Howarth EM, Jennings PE, Hepburn DA, Atkin SL. Beneficial effects of soy phytoestrogen intake in postmenopausal women with type 2 diabetes. Diabetes Care. 2002;25:1709–14. [DOI] [PubMed] [Google Scholar]
- 10.Li Z, Hong K, Saltsman P, DeShields S, Bellman M, Thames G, Liu Y, Wang HJ, Elsashoof R, Heber D. Long-term efficacy of soy-based meal replacements vs an individualized diet plan in obese type II DM patients: relative effects on weight loss, metabolic parameters, and C-reactive protein. Eur J Clin Nutr. 2005;59:411–8. [DOI] [PubMed] [Google Scholar]
- 11.Bonacasa B, Siow RC, Mann GE. Impact of dietary soy isoflavones in pregnancy on fetal programming of endothelial function in offspring. Microcirculation. 2011;18:270–85. [DOI] [PubMed] [Google Scholar]
- 12.Soucy NV, Parkinson HD, Sochaski MA, Borghoff SJ. Kinetics of genistein and its conjugated metabolites in pregnant Sprague-Dawley rats following single and repeated genistein administration. Toxicol Sci. 2006; 90:230–40. [DOI] [PubMed] [Google Scholar]
- 13.Wagner JD, Jorgensen MJ, Cline JM, Lees CJ, Franke AA, Zhang L, Ayers MR, Schultz C, Kaplan JR. Effects of soy vs. casein protein on body weight and glycemic control in female monkeys and their offspring. Am J Primatol. 2009;71:802–11. [DOI] [PubMed] [Google Scholar]
- 14.Dolinoy DC, Weidman JR, Waterland RA, Jirtle RL. Maternal genistein alters coat color and protects Avy mouse offspring from obesity by modifying the fetal epigenome. Environ Health Perspect. 2006;114:567–72. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Cederroth CR, Vinciguerra M, Gjinovci A, Kuhne F, Klein M, Cederroth M, Caille D, Suter M, Neumann D, James RW, et al. Dietary phytoestrogens activate AMP-activated protein kinase with improvement in lipid and glucose metabolism. Diabetes. 2008;57:1176–85. [DOI] [PubMed] [Google Scholar]
- 16.National Health and Nutrition Examination Survey Data. [cited 2013 June 1]. Available from: http://www.cdc.gov/nchs/nhanes/about_nhanes.htm.
- 17.National Health and Nutrition Examination Survey. NHANES 2001–2002. [cited 2013 June 1]. Available from: http://www.cdc.gov/nchs/nhanes/nhanes2001–2002/nhanes01_02.htm.
- 18.National Health and Nutrition Examination Survey. NHANES 2003–2004. [cited 2013 June 1]. Available from: http://www.cdc.gov/nchs/nhanes/nhanes2003–2004/nhanes03_04.htm.
- 19.National Health and Nutrition Examination Survey. NHANES 2005–2006. [cited 2013 June 1]. Available from: http://www.cdc.gov/nchs/data/nhanes/nhanes_05_06/general_data_release_doc_05_06.pdf.
- 20.National Health and Nutrition Examination Survey. NHANES 2007–2008. [cited 2013 June 1]. Available from: http://www.cdc.gov/nchs/nhanes/nhanes2007–2008/generaldoc_e.htm.
- 21.National Health and Nutrition Examination Survey Data Division of Laboratory Sciences Laboratory Protocol. [cited 2013 June 1]. Available from: http://www.cdc.gov/nchs/data/nhanes/nhanes_03_04/l06phy_c_met.pdf.
- 22.Wallace TM, Levy JC, Matthews DR. Use and abuse of HOMA modeling. Diabetes Care. 2004;27:1487–95. [DOI] [PubMed] [Google Scholar]
- 23.National Health and Nutrition Examination Survey 2003–2004 data documentation, codebook, and frequencies. [cited 2013 June 1]. Available from: http://www.cdc.gov/nchs/nhanes/nhanes2003–2004/FFQDC_C.htm.
- 24.Balk E, Chung M, Chew P, Ip S, Raman G, Kuplenick B. Effects of soy on health outcomes. Evid Rep Technol Assess (Summ). 2005;126:1–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Anderson JW, Johnstone BM, Cook-Newell ME. Meta-analysis of the effects of soy protein intake on serum lipids. N Engl J Med. 1995;333:276–82. [DOI] [PubMed] [Google Scholar]
- 26.Reynolds K, Chin A, Lees KA, Nguyen A, Bujnowski D, He J. A meta-analysis of the effect of soy protein supplementation on serum lipids. Am J Cardiol. 2006;98:633–40. [DOI] [PubMed] [Google Scholar]
- 27.Zhan S, Ho SC. Meta-analysis of the effects of soy protein containing isoflavones on the lipid profile. Am J Clin Nutr. 2005;81:397–408. [DOI] [PubMed] [Google Scholar]
- 28.Zhuo XG, Melby MK, Watanabe S. Soy isoflavone intake lowers serum LDL cholesterol: a meta-analysis of 8 randomized controlled trials in humans. J Nutr. 2004;134:2395–400. [DOI] [PubMed] [Google Scholar]
- 29.Taku K, Umegaki K, Sato Y, Taki Y, Endoh K, Watanabe S. Soy isoflavones lower serum total and LDL cholesterol in humans: a meta-analysis of 11 randomized controlled trials. Am J Clin Nutr. 2007;85:1148–56. [DOI] [PubMed] [Google Scholar]
- 30.Ricci E, Cipriani S, Chiaffarino F, Malvezzi M, Parazzini F. 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;17:1080–6. [DOI] [PubMed] [Google Scholar]
- 31.Sattar N, Greer IA, Louden J, Lindsay G, McConnell M, Shepherd J, Packard CJ. Lipoprotein subfraction changes in normal pregnancy: threshold effect of plasma triglyceride on appearance of small, dense low density lipoprotein. J Clin Endocrinol Metab. 1997;82:2483–91. [DOI] [PubMed] [Google Scholar]
- 32.Qureshi IA, Xi XR, Limbu YR, Bin HY, Chen MI. Hyperlipidaemia during normal pregnancy, parturition and lactation. Ann Acad Med Singapore. 1999;28:217–21. [PubMed] [Google Scholar]
- 33.Fåhraeus L, Larsson-Cohn U, Wallentin L. Plasma lipoproteins including high density lipoprotein subfractions during normal pregnancy. Obstet Gynecol. 1985;66:468–72. [PubMed] [Google Scholar]
- 34.Kühl C. Etiology and pathogenesis of gestational diabetes. Diabetes Care. 1998;21(Suppl. 2):B19–26. [PubMed] [Google Scholar]
- 35.Karcaaltincaba D, Kandemir O, Yalvac S, Guvendag-Guven S, Haberal A. Prevalence of gestational diabetes mellitus and gestational impaired glucose tolerance in pregnant women evaluated by National Diabetes Data Group and Carpenter and Coustan criteria. Int J Gynaecol Obstet. 2009;106:246–9. [DOI] [PubMed] [Google Scholar]
- 36.Linne Y. Effects of obesity on women’s reproduction and complications during pregnancy. Obesity Rev. 2004; 5:137–43. [DOI] [PubMed] [Google Scholar]
- 37.Godsland IF. Oestrogens and insulin secretion. Diabetologia. 2005;48:2213–20. [DOI] [PubMed] [Google Scholar]
- 38.Liu D, Zhen W, Yang Z, Carter JD, Si H, Reynolds KA. Genistein acutely stimulates insulin secretion in pancreatic beta-cells through a cAMP-dependent protein kinase pathway. Diabetes. 2006;55:1043–50. [DOI] [PubMed] [Google Scholar]
- 39.Ohno T, Kato N, Ishii C, Shimizu M, Ito Y, Tomono S, Jawazu S. Genistein augments cyclic adenosine 3′5′-monophosphate(cAMP) accumulation and insulin release in MIN6 cells. Endocr Res. 1993;19:273–85. [DOI] [PubMed] [Google Scholar]
- 40.Sorenson RL, Brelje TC, Roth C. Effect of tyrosine kinase inhibitors on islets of Langerhans: evidence for tyrosine kinases in the regulation of insulin secretion. Endocrinology. 1994;134:1975–8. [DOI] [PubMed] [Google Scholar]
- 41.Choi MS, Jung UJ, Yeo J, Kim MJ, Lee MK. 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. [DOI] [PubMed] [Google Scholar]
- 42.Mezei O, Banz WJ, Steger RW, Peluso MR, Winters TA, Shay N. Soy isoflavones exert antidiabetic and hypolipidemic effects through the PPAR pathways in obese Zucker rats and murine RAW 264.7 cells. J Nutr. 2003;133:1238–43. [DOI] [PubMed] [Google Scholar]
- 43.Bhathena SJ, Velasquez MT. Beneficial role of dietary phytoestrogens in obesity and diabetes. Am J Clin Nutr. 2002;76:1191–201. [DOI] [PubMed] [Google Scholar]
- 44.Frankenfeld CL. Dairy consumption is a significant correlate of urinary equol concentration in a representative sample of US adults. Am J Clin Nutr. 2011;93:1109–16. [DOI] [PubMed] [Google Scholar]
- 45.Wu X, Cai H, Gao YT, Dai Q, Li H, Cai Q, Yang G, Franke AA, Zheng W, Shu XO. Correlations of urinary phytoestrogen excretion with lifestyle factors and dietary intakes among middle-aged and elderly Chinese women. Int J Mol Epidemiol Genet. 2012; 3:18–29. [PMC free article] [PubMed] [Google Scholar]
- 46.Decroos K, Vanhemmens S, Cattoir S, Boon N, Verstraete W. Isolation and characterisation of an equol-producing mixed microbial culture from a human faecal sample and its activity under gastrointestinal conditions. Arch Microbiol. 2005;183:45–55. [DOI] [PubMed] [Google Scholar]
- 47.Hedlund TE, Johannes WU, Miller GJ. Soy isoflavonoid equol modulates the growth of benign and malignant prostatic epithelial cells in vitro. Prostate. 2003;54:68–78. [DOI] [PubMed] [Google Scholar]
- 48.Setchell KD, Cole SJ. Method of defining equol-producer status and its frequency among vegetarians. J Nutr. 2006;136:2188–93. [DOI] [PubMed] [Google Scholar]
- 49.Akaza H, Miyanaga N, Takashima N, Naito S, Hirao Y, Tsukamoto T, Fujioka T, Mori M, Kin WJ, Song JM, et al. Comparisons of percent equol producers between prostate cancer patients and controls: case-controlled studies of isoflavones in Japanese, Korean and American residents. Jpn J Clin Oncol. 2004;34:86–9. [DOI] [PubMed] [Google Scholar]
- 50.Setchell KD, Brown NM, Lydeking-Olsen E. The clinical importance of the metabolite equol-a clue to the effectiveness of soy and its isoflavones. J Nutr. 2002;132:3577–84. [DOI] [PubMed] [Google Scholar]
- 51.Chun OK, Chung SJ, Song WO. Urinary isoflavones and their metabolites validate the dietary isoflavone intakes in US adults. J Am Diet Assoc. 2009;109:245–54. [DOI] [PubMed] [Google Scholar]
- 52.Bolca S, Possemiers S, Herregat A, Huybrechts I, Heyerick A, De Vriese S, Verbruggen M, Depypere H, De Keukeleire D, Bracke M, et al. Microbial and dietary factors are associated with the equol producer phenotype in healthy postmenopausal women. J Nutr. 2007;137:2242–6. [DOI] [PubMed] [Google Scholar]
- 53.Kelly GE, Joannou GE, Reeder AY, Nelson C, Waring MA. The variable metabolic response to dietary isoflavones in humans. Proc Soc Exp Biol Med. 1995;208:40–3. [DOI] [PubMed] [Google Scholar]
- 54.Rybak ME, Sternberg MR, Pfeiffer CM. Sociodemographic and lifestyle variables are compound- and class-specific correlates of urine phytoestrogen concentrations in the U.S. population. J Nutr. 2013;143:986S–94S. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 55.Frankenfeld CL, Patterson RE, Horner NK, Neuhouser ML, Skor HE, Kalhorn TF, Howald WN, Lampe JW. Validation of a soy food-frequency questionnaire and evaluation of correlates of plasma isoflavone concentrations in postmenopausal women. Am J Clin Nutr. 2003;77:674–80. [DOI] [PubMed] [Google Scholar]
- 56.Valentín-Blasini L, Blount BC, Caudill SP, Needham LL. Urinary and serum concentrations of seven phytoestrogens in a human reference population subset. J Expo Anal Environ Epidemiol. 2003;13:276–82. [DOI] [PubMed] [Google Scholar]
- 57.Lee SA, Wen W, Xiang YB, Barnes S, Liu D, Cai Q, Zheng W, Shu XO. Assessment of dietary isoflavone intake among middle-aged Chinese men. J Nutr. 2007;137:1011–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 58.Atkinson C, Skor HE, Fitzgibbons ED, Scholes D, Chen C, Wahala K, Schwartz SM, Lampe JW. Overnight urinary isoflavone excretion in a population of women living in the United States, and its relationship to isoflavone intake. Cancer Epidemiol Biomarkers Preven. 2002;11:253–60. [PubMed] [Google Scholar]