Soybean and soy isoflavone intake indicate a positive change in bone mineral density for 2 years in young Korean women

Available online at www.sciencedirect.com

Nutrition Research 28 (2008) 25 – 30 www.elsevier.com/locate/nutres

Soybean and soy isoflavone intake indicate a positive change in bone mineral density for 2 years in young Korean women YoonJu Song a , Hee Young Paik a , Hyojee Joung b,⁎ a

Department of Food and Nutrition, Seoul National University, Seoul 151-742, Korea Graduate School of Public Health, Seoul National University, 28 Yungun-dong Chongro-gu, Seoul 110-799, Korea Received 30 April 2007; revised 27 September 2007; accepted 9 November 2007

b

Abstract Many studies have shown that soybean and isoflavones have a favorable effect on bone mass for postmenopausal women, but few data are available on young Asian women. To investigate the effect of soybean and isoflavone intake on bone mineral density (BMD) and its change among young Korean women over 2 years, we conducted a longitudinal study for 34 women. The BMD was measured 3 times with 1-year intervals by dual x-ray absorptiometry at the lumbar spine and femur (neck, Ward's triangle [WT], and trochanter). Dietary intake was assessed up to 8 times by 24-hour recall with average 4-month interval. During the study period, BMD increased significantly for lumber spine and WT (2.5% and 5.2%). The average daily intake of soybeans and isoflavones was 39 g and 8 mg, respectively. Soybean intake and total isoflavone intake had positive correlation on femoral neck (FN) and WT. By longitudinal mixed-model regression analysis, BMD increased 0.26% per 1 mg of isoflavone intake per year in the FN and 0.31% for WT (P = .05 and .008). In conclusion, soybean and isoflavone intake have a positive effect on the change of BMD on the FN and WT among young Korean women. Because soybean and isoflavone intakes could be confounded by other nutrients, the positive effects of isoflavones on bone should be further investigated. © 2008 Elsevier Inc. All rights reserved. Keywords:

Bone mineral density change; Femoral neck; Wards' triangle; Isoflavones; Soybeans; Young Korean women

1. Introduction An increasing number of studies have investigated the effect of soy intake on bone health. Although recent reviews acknowledge that these findings are inconclusive, they indicate that soy isoflavones may have favorable benefits on bone for Asian populations [1-3]. Eleven recent epidemiologic studies, including clinical trials and longitudinal studies, addressed the association between soybeans or isoflavone intake and bone mineral density (BMD) among Asian populations [4-14]. Eight studies reported favorable effects of isoflavone supplemen⁎ Corresponding author. Tel.: +82 2 740 8865; fax: +82 2 745 9104. E-mail address: [email protected] (H. Joung). 0271-5317/$ – see front matter © 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.nutres.2007.11.001

tation or high soy isoflavone intake on at least one bone site [4-11]. Others found some associations with bone markers [12] or no association [13,14]. Most of these studies were conducted with postmenopausal women; only 3 studies included young women. Young adulthood and adolescence are a critical period because of rapid skeletal bone mass accumulation and nearly half of the adult bone mass has accrued. It is important to attain peak bone mass and to reduce the risk of osteoporosis later in life [15]. Thus, any benefits of soy consumption on bone mass for young women should be investigated. Among 3 studies, only one looked at Chinese adolescent girls aged 14 to 16 years and found significant positive change in BMD at the hip after 1 year of calcium-fortified soy milk supplementation [8]. Other studies examined a broad range

26

Y. Song et al. / Nutrition Research 28 (2008) 25–30

of age and found some favorable effect of soy isoflavones on BMD change, but it was not possible to distinguish the effects of calcium and isoflavones [5,11]. However, the study for white young women found no effect on BMD after 1 year of soy protein supplementation [16]. A difference could be explained by the soybean intake between Asian and white people even though a number of dietary conditions such as calcium and protein might affect the association between isoflavones and bone. In Asian countries including China, Japan, and Korea, soy is part of the traditional diet. Based on the Korean National Health and Nutrition Examination Survey in 1995, the average intake of dietary isoflavones was reported to about 15 mg [17], which is higher than that of other western countries and comparable with those of Chinese and Japanese women [11,17-20]. In Korea, 3 traditional soy foods, tofu, soybean paste, and soybean sprout, contribute to more than 94% of total isoflavone intake [17]. The habitual high intake of soybeans could bring favorable benefits on bone mass for Asian populations. Thus, we hypothesized that habitual soybeans/isoflavone intake could affect BMD or its change for young Asian women. In previous studies for Asian populations, there were no longitudinal studies that evaluated the effects of habitual isoflavone intake on BMD, specifically for young women. Therefore, the objective of this study was to assess the habitual intake of soybeans and isoflavones using multiple 24-hour recalls and to examine the association between isoflavones and BMD among healthy young Korean women aged 20 to 26. 2. Methods and materials 2.1. Subjects and study design The study participants were healthy young women in Seoul, Korea. Volunteers were recruited through fliers posted at Seoul National University, Seoul, Korea. Eligibility criteria included willingness to participate in the study protocol and understanding and signing of informed consent (ability to perform all assessments including multiple 24-hour recalls and bone measurements 3 times). Exclusion criteria included previous diagnosis of bone disease, arthritis, and any serious chronic disease such as diabetes, liver disease, hyperthyroidism, and chronic drug use. Fifty-five subjects, age range from 20 to 26 years, participated at the baseline after screening. At baseline, BMD, height, and weight were assessed by professionals at Seoul National University Hospital, and one of the participants of the 24-hour recall was interviewed by dietitians. After the baseline, bone measurements were made, each 1 year apart, for a total of 3 times, and participants completed up to eight 24-hour recalls approximately every 4 months. Only 34 subjects completed all 3 BMD measurements and were considered in the final sample size for this study (we focused on the slope for BMD). The study protocol was

approved by the Human Research Committee at the College of Human Ecology in Seoul National University. 2.2. Bone mineral density and anthropometric measurements Bone mineral density was measured by dual energy x-ray absorptiometry (LUNAR expert; LUNAR Co., Madison, Wis) at the following sites: lumbar spine (L2-L4) and 3 different sites of the hip—femoral neck (FN), Ward's triangle (WT), and femoral trochanter (FT), by the Department of Radiology in Seoul National Hospital. Outcomes are reported as percentage change from baseline. Height and body weight were measured with an upright balance scale (Jenix; Dongshin Co, Seoul, Korea). 2.3. Dietary assessment Twenty-four hour recalls were conducted for each subject at 4-month intervals across the study period to assess average intake. The maximum number of recalls for the subjects was 8, and the average number of recalls for the subjects was 6.6. The 24-hour recalls were converted to daily nutrient intake using software developed in our laboratory with Nutrient Database [21,22]. The recalls were averaged to obtain a single daily nutrient report across the study. To calculate the intake of soybeans and their products, we grouped all food items into 18 food groups by classification using the Korean Nutrient Database. The soybean group included various types of soybean, soybean curd (tofu), soybean milk, and small red beans. The average intake of soybean (g) and the percentage of energy from soybeans were calculated based on daily energy. To calculate the frequency of soybean food intakes, we defined “the soybean consumption rate” as the number of days where soybeans were consumed out of the total number of recalls conducted. Because soybeans are usually served in side dishes in Korea, no minimum amount consumed per day was required to be included in our calculation. Consumption of soybeans was dichotomized into any consumption = 1 or no consumption = 0. The soybean consumption rate is the percentage of days with any consumption of soybeans divided by the total number of recalls conducted for that participant. Isoflavone intake and average energy from isoflavones were compiled using the isoflavone database, which was established by Park et al [22] for food items in the Korean Nutrient Database. This database was established only with soybean food items. Isoflavone intake in this study is defined as “soy isoflavone” intake. 2.4. Statistical methods All statistical analyses were performed using Statistical Analysis System (release 9.1; SAS Inc, Cary, NC). Descriptive values including demographic variables and soybean intake were presented with mean ± SD and range. Spearman correlation analysis determined relationships

Y. Song et al. / Nutrition Research 28 (2008) 25–30 Table 1 General characteristics of subjects participating in the study Mean a ± SD Age (y) Menarche (y) Height (cm) Weight (kg) BMI (kg/m2)

3. Results Range

22.1 ± 1.2 12.6 ± 1.2 159.8 ± 4.9 51.4 ± 6.1 20.1 ± 2.1

27

20-26 10-15 152-168 42.4-63.0 16.4-25.1 BMD (g/cm2)

Mean ± SD

Range

At baseline Lumbar spine FN WT FT

1.121 ± 0.137 0.965 ± 0.098 0.891 ± 0.124 0.776 ± 0.111

0.942-1.479 0.720-1.251 0.587-1.212 0.556-1.201

After 1 y Lumbar spine FN ⁎⁎ WT FT ⁎⁎

1.132 ± 0.143 0.930 ± 0.101 0.890 ± 0.125 0.754 ± 0.100

0.937-1.563 0.684-1.171 0.625-1.199 0.553-1.137

After 2 y Lumbar spine ⁎ FN WT ⁎⁎ FT

1.149 ± 0.141 0.980 ± 0.109 0.936 ± 0.129 0.759 ± 0.114

0.950-1.579 0.704-1.310 0.643-1.272 0.528-1.169

Sample size, n = 34. a Mean values were significantly different between baseline and each year by paired t test. ⁎ P b .05. ⁎⁎ P b .001.

between soybean/isoflavone intake and BMD change because isoflavones and BMD were not normally distributed. For analyses of repeated measures of BMD, we used a mixed-effects regression model with a random intercept and compound symmetry covariance matrix (Proc Mixed) [23,24]. This method is a generalized form of linear regression analysis that allows for repeated measures on each subject while accounting for the considerable variation across subjects in overall average BMD and BMD change over time. The main outcome was individual BMD measurements and covariates, which were treated as main effects and time-dependent effects. In this model, covariates were age at baseline, menarche, body mass index at baseline, and average total isoflavone intake, and all covariates were included with both forms such as subject-level intercepts and slope for time. Regarding the skew in the distributions of BMD, we used log transformation for BMD. Thus, to estimate trends for isoflavone intake on BMD change, we used the regression estimate β* for isoflavones as slope for time and calculated the corresponding percentage change in BMD over time by the monotonic transformation 100× (eβ* - 1) [23,24]. All levels of significance were set at P b .05 unless otherwise stated.

3.1. Bone mineral density over time Demographics and BMD measurements for study participants are shown in Table 1. The mean age for study subjects was 22.1 years (SD = 1.2), and the mean age at menarche was 12.6 years (SD = 1.2). Mean BMD at the lumbar spine was 1.121, 1.132, and 1.149 g/cm2 at each year. Average lumbar BMD increased slightly each year and was significantly higher between baseline and 2-year follow-up (P = .04). Femoral BMD at all sites—FN, WT, and FT—showed similar trends to decrease at the 1-year follow-up and to increase at the 2-year followup. The percentage change of BMD at the 2-year follow-up was 2.5% for lumbar spine, 1.6% for FN, 5.2% for WT, and −2.2% for FT. The changes in BMD over 2 years were significantly different from baseline for lumbar spine (P = .04) and WT (P = .0001), but not for FN and trochanter. 3.2. Nutrient intake over time The nutrient and food intake are presented in Table 2. The average daily intake of energy was 7550 kJ, protein was 65 g, and fat was 48 g. The percentage of energy from carbohydrate was 61% and from fat was 24%. The average daily calcium intake was 490 mg, and the average daily iron intake was 12 mg. The average daily intake of food in the milk group was 194 g, and its percentage of daily energy intake was 5.7%. The average daily intake of the soybean group was 39 g with a wide range from 0 to 137 g. The contribution of daily energy intake from the soybean group was 1.2%. The

Table 2 The average daily dietary intake of women during the study period

Energy (kJ) Protein (g) Fat (g) Calcium (mg) Phosphorous (g) Iron (mg) Vitamin C (mg) Milk and their products Intake (g) Intake (% energy) Soybeans Intake (g) Intake (% energy) Consumption rate a Isoflavonoids Daidzein (mg/d) Genistein (mg/d) Glycitein (mg/d) Total isoflavone (mg/d)

Mean ± SD

Range

7550 ± 242 65.4 ± 13.5 48.4 ± 9.6 489 ± 138 915 ± 177 11.6 ± 2.5 83.4 ± 28

5497-9172 44.1-91.9 33.1-76.3 264-858 613-1322 7.8-19.4 39.9-151

194 ± 93 5.66 ± 2.91

0-444 0-13.9

39.1 ± 30.8 1.22 ± 0.95 0.46 ± 0.25

0-137 0-3.69 0-1

3.41 ± 2.30 3.47 ± 2.11 1.13 ± 0.60 7.96 ± 4.52

0.55-11.9 0.59-8.23 0.19-3.06 1.45-20.23

Sample size, n = 34. a Soybean consumption rate = no. of days reported to consume soybean among total days of recall per person.

28

Y. Song et al. / Nutrition Research 28 (2008) 25–30

Table 3 Correlation analysis between soybeans/isoflavone intake and BMD change in women BMD change (% for 2 y) Lumbar spine Soybeans intake Intake (g/d) Intake (% energy/d) Consumption rate a Isoflavone intake Daidzein (mg) Genistein (mg) Glycitein (mg) Total isoflavones (mg)

FN

WT

FT

0.028 0.060 0.152

0.192 0.328 ⁎ 0.371 ⁎

0.468 ⁎⁎ 0.506 ⁎⁎ 0.433 ⁎

−0.015 −0.028 −0.064

−0.004 0.024 0.150 0.032

0.347 ⁎ 0.358 ⁎ −0.056 0.337 ⁎

0.510 ⁎⁎ 0.519 ⁎ 0.163 0.506 ⁎⁎

−0.034 −0.018 0.298 −0.002

Sample size, n = 34. a Soybean consumption rate = no. of days reported to consume soybean among total days of recall per person. ⁎ Spearman correlation coefficients, P ≤ .05. ⁎⁎ Spearman correlation coefficients, P b .01.

soybean consumption rate was 0.46, which means that participants consumed soybeans almost every other day. Analyzing the specific components of isoflavones revealed that the average daily intake was 3.4 mg for daidzein, 3.5 mg for genistein, 1.1 mg for glycitein, and 8.0 mg for total isoflavones. 3.3. Effect of isoflavone on BMD Table 3 presents the association between soybeans and isoflavone intake with the change of BMD over 2 years. A significant and positive relationship was observed between the average daily percentage of energy from soybeans and BMD change for FN (r = 0.33, P = .05) and WT (r = 0.51, P = .002). Similar associations with isoflavone intake were evidenced with FN (r = 0.34, P = .05) and WT BMD (r = 0.51, P = .002), except for the glycitein subsite. The mixed-model analyses are shown in Table 4 for the association between isoflavones and change in BMD at 2-year follow-up. For every 1-mg increase of isoflavones, BMD significantly increased at a rate of 0.26% per year for FN and 0.31% per year for WT after adjusting for age, menarche, body mass index (BMI), and baseline BMD. 4. Discussion In this study, we hypothesized that soybeans and isoflavone intakes were associated with BMD or its change among young Korean women. Cross-sectional associations of soybeans and isoflavone intake with BMD were not observed at any year. However, we found a significant positive association between soybeans and isoflavone intakes and the change of BMD over 2 years especially at the FN and WT. Our study was not an intervention study but showed increased percentage of BMD with high soy isoflavone intake consistent with other longitudinal studies [5,6].

The average daily intake of soy isoflavones in this study was 8 mg, which is lower than that in other studies that reported a beneficial effect. However, our study is the first study to assess soy isoflavones with multiple 24-hour recalls over 2 years; most other studies used Food Frequency Questionnaire at a single point. Our isoflavone intake measures thus represent average habitual isoflavone intake during the study. Our average is similar with the study among Chinese women by Ho et al [7]. Their study showed that habitual high soy isoflavone intake had a positive effect on the slope of spine BMD for Chinese women over 3 years, which is consistent with our findings. In addition to average daily soy intake, the soybean consumption rate showed a significantly positive correlation with change in BMD over 2 years at WT. Our participants consumed soybeans approximately every other day. The combination of high number of days of consumption as well as the amount of soybeans would contribute to favorable effects on bone. According to its estrogenic action, isoflavones might be effective when estrogen is deficient or when estrogen is not sufficient [1]. A study in white, young, healthy women with normal menstrual cycles, which is in the same age range with that of our study, reported that soy protein–supplemented group had no effect on bone compared with the control after 1 year [16]. The authors made the argument that the estrogen status of white young women is sufficient compared with older menopausal subjects with ovarian hormone deficiency. However, considering the fact that the Asian population has a lower BMD and body size [25,26], estrogen status for Asian young women would not be as sufficient as white young women even though they are healthy and with normal menstrual cycle. Two Korean studies supported this explanation. First, isoflavone supplementation over 3 menstrual cycles improved BMD for young college women who were underweight (mean BMI, 18.2), but had no change among women with normal weight (mean BMI, 20.8) where there was no difference between the 2 groups regarding the onset of menarche, menstrual cycle, and regularity [27]. Second, soy milk supplementation for 10 weeks increased BMD in lumber spine and FN in underweight college women (mean BMI, 18.8) [28]. As the perception of beauty in Korea is to be thin Table 4 Repeated measures analysis for association of isoflavone intake with BMD over time in women BMD change (% increase/y) a

Isoflavone intake (mg) P

Lumbar spine

FN

WT

FT

0.01 .90

0.26 .05

0.31 .008

−0.01 .88

Mixed-effects regression model was used, and the model included a time variable, main effects of covariates, and corresponding interactions with the time variable (see Statistical methods section). Covariates were age, menarche, body mass index, BMD at baseline, and isoflavone intake. a Percentage increase in BMD was associated with 1 mg of isoflavone intake.

Y. Song et al. / Nutrition Research 28 (2008) 25–30

and slim, the underweight would be the major concern along with low BMD, especially for young women in Korea [28]. Our finding showed only favorable effect on bone change. Soy isoflavones may affect young women who are low in BMD at baseline and help to improve their BMD. In addition, habitual isoflavone intake among the Asian population would be easy to comply with and to play a role in estrogenic action. The favorable effects of isoflavones and soybeans on BMD change were found only in some femur sites, specifically WT and FN. This may be because of the different composition of each skeletal site. Bone is composed of cortical and trabecular bone that plays different roles in bone metabolism. Regarding calcium balance, trabecular bone is more sensitive to hormones that handle day-to-day deposits and removal of calcium [29]. Ward's triangle at the femur is one of the sites that has abundant trabecular bone, thus, at which BMD change would be more affected by environmental factors such as isoflavone intake. Besides, it has been reported that there are skeletal site-dependent responses of BMD to environmental effects such as dietary intervention, and different responses between white and Asian women occur [30-32]. However, this needs further research to explain which skeletal sites acquire more benefits from isoflavones among the Asian population. There were several limitations in this study. The major challenge was the isoflavone database. We used the isoflavone database [22] where isoflavones were compiled by the US Department of Agriculture systematical evaluation system for 142 soybean items, which Koreans usually consume. Its advantage is that it incorporates the variety of soybeans, the crop year, and the location, all of which affect the isoflavone content of the soybeans because it used all domestic data. Difficulties arose because it cannot be directly compared with other studies. Another limitation was that, because soybean intake was so low in this study, other food or nutrients might be responsible for the change in BMD over 2 years. Although calcium and protein intake were not associated with change of BMD in this study, those nutrients might interact with isoflavones in synergistic ways for bone metabolism. There is, however, evidence in clinical trials that isoflavones have the potential to both stimulate bone formation and inhibit bone resorption, although the mechanism of how isoflavones mediate bone metabolism is not clear [1,2]. Finally, we had a small sample size, and our results might not represent the entire spectrum of BMD as well as soy isoflavone intake for young Korean women. However, our study period was 2 years, which is the sufficient time to assess the effects. Because bone is a slowly responding organ, longer periods such as 2 to 3 years have been suggested to best predict favorable effects of any factors on bone [1,2]. Future studies should examine the effects of soybeans and isoflavone intake on BMD change in a larger population. In conclusion, soy isoflavone intake had a positive effect on the change in BMD at FN and WT for young Korean

29

women. This study is the first study to report the benefits of isoflavone intake with BMD change using multiple 24-hour recalls for young Asian women. Our finding suggests that habitual isoflavone intake among the Asian population would be beneficial on BMD change especially for young women with low BMD. Habitual soybean intake was low in this study though, and other dietary components also affect bone metabolism. Therefore, future studies should address the interaction of isoflavones with other dietary components. In this case, the mechanism of how isoflavones affect bone metabolism and the use of bone biomarker would result in a better understanding in young Asian women. References [1] Weaver CM, Cheong JMK. Soy isoflavones and bone health: the relationship is still unclear. J Nutr 2005;135:1243-7. [2] Messina M, Ho S, Alekel DL. Skeletal benefits of soy isoflavones: a review of the clinical trial and epidemiologic data. Curr Opin Clin Nutr Metab Care 2004;7:649-58. [3] Chen X, Anderson JJ. Isoflavones and bone: animal and human evidence of efficacy. J Musculoskelet Neuronal Interact 2002;2:352-9. [4] Chen YM, Ho SC, Lam SSH, Ho SSS, Woo JLF. Soy isoflavones have a favorable effect on bone loss in Chinese postmenopausal women with lower bone mass: a double-blind, randomized, controlled trial. J Clin Endocrinol Metab 2003;88:4740-7. [5] Ikeda Y, Iki M, Morita A, Kajita E, Kagamimori S, Kagawa Y, et al. Intake of fermented soybeans, Natto, is associated with reduced bone loss in postmenopausal women: Japanese population-based osteoporosis (JPOS) study. J Nutr 2006;136:1323-8. [6] Huang H, Yang H, Yang H, Yang T, Shieh M, Huang S. One-year soy isoflavone supplementation prevents early postmenopausal bone loss but without a dose-dependent effect. J Nutr Biochem 2006;17:509-17. [7] Ho SC, Woo J, Lam S, Chen Y, Sham A, Lau J. Soy protein consumption and bone mass in early postmenopausal Chinese women. Osteoporos Int 2003;14:835-42. [8] Ho SC, Guldan GS, Woo J, Yu R, Tse MM, Sham A, et al. A prospective study of the effects of 1-year calcium-fortified soy milk supplementation on dietary calcium intake and bone health in Chinese adolescent girls aged 14 to 16. Osteoporos Int 2005;16:1907-16. [9] Mei J, Yeung SSC, Kung AWC. High dietary phytoestrogen intake is associated with higher bone mineral density in postmenopausal but not premenopausal women. J Clin Endocrinol Metab 2001;86:5217-21. [10] Tsuchida K, Mizushima S, Toba M, Soda K. Dietary soybeans intake and bone mineral density among 995 middle-aged women in Yokohama. J Epidemiol 1999;9:14-9. [11] Ho SC, Chan SG, Yi Q, Wong E, Leung PC. Soy intake and the maintenance of peak bone mass in Hong Kong Chinese Women. J Bone Miner Res 2001;16:1363-9. [12] Uesugi T, Toda T, Okuhira T, Chen J. Evidence of estrogenic effect by the three-month-intervention of isoflavone on vaginal maturation and bone metabolism in early postmenopausal women. Endocr J 2003;50:613-9. [13] Hsu CS, Shen WW, Hsueh YM, Yeh SL. Soy isoflavone supplementation in postmenopausal women. Effects on plasma lipids, antioxidant enzyme activities and bone density. J Reprod Med 2001;46:221-6. [14] Nagata C, Shimizu H, Takami R, Hayashi M, Takeda N, Yasuda K. Soy product intake and serum isoflavonoid and estradiol concentration in relation to bone mineral density in postmenopausal Japanese women. Osteoporos Int 2002;13:200-4. [15] Weaver CM. Adolescence: the period of dramatic bone growth. Endocrine 2002;17:43-8. [16] Anderson JJB, Chen X, Boass A, Symons M, Kohlmeier M, Renner JB, et al. Soy isoflavones: no effects on bone mineral content and bone

30

[17]

[18]

[19]

[20]

[21] [22]

[23] [24]

Y. Song et al. / Nutrition Research 28 (2008) 25–30 mineral density in health, menstruation young adult women after one year. J Am Coll Nutr 2002;21:388-93. Kim J, Kwon C. Estimated dietary isoflavone intake of Korean population based on National Nutrition Survey. Nutr Res 2001;21:947-53. Maskarinec G, Singh S, Meng L, Franke AA. Dietary soy intake and urinary isoflavone excretion among women from a multiethnic population. Cancer Epidemiol Biomarkers Prev 1998;7:613-9. Greendale GA, FitzGerald G, Huang M, Sternfeld B, Gold E, Seeman T, et al. Dietary soy isoflavones and bone mineral density: results from the study of women's health across the nation. Am J Epidemiol 2002;155:746-54. Van Erp-Baart MA, Brants HA, Kiely M, Mulligan A, Turrini A, Sermoneta C, et al. Isoflavone intake in four different European countries: the VENUS approach. Br J Nutr 2003;89:S25-S30. Korean Nutrition Society. Dietary reference intakes for Koreans. Seoul: Kookjin Publishing Co; 2005 [in Korean]. Park MK, Song YJ, Joung HJ, Li S, Paik HY. Establishment of an isoflavone database for usual Korean foods and evaluation of isoflavone intake among Korean children. Asia Pac J Clin Nutr 2007;16:129-39. Diggle PJ, Heagerty P, Liang K-Y, Zeger S. Analysis of longitudinal data. 2nd ed. Oxford, UK: Oxford University Press; 2002. SAS Institutes Inc. Base SAS 9.1.3 Procedures Guide, vols 1-4. Base SAS 9.1.3 (TS1M3) edition. Cary (NC): SAS publishing; 2004.

[25] Ho SC. Body measurement, bone mass, and fractures. Does the east differ from the west? Clin Orthop Relat Res 1996;323:75-80. [26] Lau EM, Lynn H, Woo J, Melton LJ. Areal and volumetric bone density in Hong Kong Chinese: a comparison with Caucasians living in the United States. Osteoporos Int 2003;14:583-8. [27] Baek SK, Sung CJ. A study of soy isoflavone supplementation effect on bone mineral density and bone metabolism markers in female college students with low bone mass. Korean J Nutr 2003;36:154-66. [28] Sung CJ, Kim SY, Kim MH, Kim EY. The effect of isoflavone supplementation by soymilk on bone mineral density in underweight college women. Korean J Nutr 2003;36:470-5. [29] Rolfes SR, Pinna K, Whitney E. Understanding normal and clinical nutrition. 7th ed. Stamford, CT: Thomson Wadsworth; 2006. p. 428-35. [30] Yang TL, Zhao LJ, Liu YJ, Liu JF, Recker RR, Deng HW. Genetic and environmental correlations of bone mineral density at different skeletal sites in females and males. Calcif Tissue Int 2006;78:212-7. [31] Moschonis G, Manios Y. Skeletal site-dependent response of bone mineral density and quantitative ultrasound parameters following a 12month dietary intervention using dairy products fortified with calcium and vitamin D: the Postmenopausal Health Study. Br J Nutr 2006;96:1140-8. [32] Wu XP, Liao EY, Huang G, Dai RC, Zhang H. A comparison study of the reference curves of bone mineral density at different skeletal sites in native Chinese, Japanese, and American Caucasian. Calcif Tissue Int 2003;73:122-32.