European Journal of Clinical Nutrition (2003) 57, 324–327 ß 2003 Nature Publishing Group All rights reserved 0954–3007/03 $25.00 www.nature.com/ejcn ORIGINAL COMMUNICATION A randomized controlled trial of phytoestrogen supplementation, growth and bone turnover in adolescent males

G Jones1*, T Dwyer1, K Hynes1, FS Dalais2, V Parameswaran3 and TM Greenaway3

1Menzies Centre for Population Health Research, Hobart, Tasmania, Australia; 2Department of Epidemiology and Preventive Medicine, Monash University, The Alfred Hospital, Prahran, Victoria, Australia; and 3Diabetes and Endocrine Services, Royal Hobart Hospital, Hobart, Tasmania

Objective: To assess the effect of phytoestrogens on bone turnover and growth in adolescent boys. Design: Randomized double-blind placebo-controlled trial. Setting: Single school in northwest Tasmania. Participants: Adolescent boys (treatment n ¼ 69, placebo n ¼ 59, mean age 16.8 y). Interventions: Six weeks of isoflavone supplementation (Novasoy, 50 mg daily of isoflavone equivalents). Bone turnover markers (bone specific alkaline phosphatase (BAP) and pyridinoline creatinine ratio (PYR)) were measured at baseline and follow-up. Results: Despite marked increases in urinary genistein and daidzein in the treatment arm (both P < 0.001), there were no significant differences in BAP, PYR or short-term height or weight change. This applied to both intention-to-treat and per protocol analysis. Neither was there a significant correlation between urinary genistein and daidzein levels and BAP or PYR. Conclusions: Phytoestrogen supplementation to the level of usual Japanese dietary intake has no measurable effect on bone turnover in adolescent boys. Longer-term studies of bone density may be desirable but it is unlikely that there will be a large effect in either girls or boys given the lower endogenous oestrogen levels in boys. Sponsorship: National Health and Medical Research Council of Australia, Arthritis Foundation of Australia. European Journal of Clinical Nutrition (2003) 57, 324 – 327. doi:10.1038=sj.ejcn.1601544

Keywords: phytoestrogens; trial; bone; adolescence

Introduction isoflavone intake. It is estimated that Japanese intake is Fractures due to osteoporosis are a significant public health around 40 – 50 mg of isoflavone per day (Kimira et al, 1998; problem. There are well-documented differences in fracture Somekawa et al, 2001), while Western intake is minimal. incidence between Asian and Caucasian populations. The Studies in rats have suggested that both isoflavones (Fanti reason for this is not fully understood but may involve hip et al, 1998; Ishida et al, 1998; Picherit et al, 2001) and axis length variations (Wang et al, 1997) or phytoestrogen or coumestrol (Draper et al, 1997) can prevent bone loss in the ovariectomy model, most likely through prevention of bone resorption (Ishida et al, 1998; Tsutsumi, 1995). Human *Correspondence: G Jones, Menzies Centre for Population Health studies have provided conflicting results on the effect of Research, GPO Box 252-23, Hobart, Tasmania 7000, Australia. these agents both on hot flushes (Dalais et al, 1998; Upmalis E-mail: [email protected] et al, 2000) and a variable, and possibly dose-dependent, Guarantor: G Jones. effect on prevention of bone loss (Alekel et al, 2000; Dalais Contributors: GJ was responsible for study conception, data analysis and overall manuscript composition. TD and TMG were responsible et al, 1998; Potter et al, 1998), but uncertain effects on bone for study conception and manuscript review. KH was responsible for turnover (Wangen et al, 2000). Fracture risk in later life is also data collection and manuscript review. FS was responsible for affected by peak bone mass. However, there have been no phytoestrogen assays and manuscript review. VP was responsible for studies of the effect of phytoestrogens in children or adoles- bone assays and manuscript review. Received 1 November 2001; revised 15 May 2002; cents. Furthermore, males have lower estrogen levels than accepted 28 May 2002 females and, in adult males, estrogen levels appear to be a Phytoestrogens and bone turnover in boys G Jones et al 325 stronger determinant of bone density than androgens was assessed by the bone-specific alkaline phosphatase (Falahati-Nini et al, 2000). The aim of this study was to immunoassay (Alkphase B, Metra Biosystems). It is highly assess the effect of 6 weeks of isoflavone supplementation specific for bone with a long half-life and responds rapidly on bone turnover markers in adolescent boys. to antiresorptive therapy in adults (Ettinger et al, 1999; Schnitzer et al, 2000). The coefficient of variation (CV) in this study was 5.6%. Bone resorption was assessed by urinary deoxypyridinoline creatinine ratio measured by competitive Methods immunoassay on a 20 ml aliquot taken from the urine The study was conducted in Burnie, Tasmania (latitude 42S) collection, protected from light by a black bag and stored at a single school (Hellyer College) between 5 July and 27 at 720C prior to analysis (Pyrilinks-D, Metra Biosystems). August 1999. All male students aged 16 – 18 y were eligible to The CV in this study was 6%. Urinary creatinine excretion participate and were invited to participate through an infor- was also assessed on a first void morning urine sample. mation sheet sent to the home address and distributed Aliquots from each sample were diluted with neutral through the school’s daily newsletter. Of the 297 boys pH buffer and analysed by the Ektachem method. The CV eligible, 136 volunteered and provided informed consent. was 5.4%. The study was granted ethical approval by the University of Tasmania Human Research Ethics Committee (Project no. H0003116). Statistics Subjects were randomly assigned to receive Novasoy Descriptive variables are listed as means or percentages. (Archer Daniels Midland, 50 mg in total isoflavones) or Unpaired t-tests or Mann – Whitney U-tests (where appropri- identical placebo taken as tablets (round, 11 mm diameter). ate) were utilized to compare baseline variable and change in The randomization was determined by the manufacturer short-term growth, bone turnover markers and urinary phy- using computer-generated random numbers and subjects toestrogen assays. For the main hypotheses, analysis utilized were allocated in the order determined by the manufacturer both an intention-to-treat approach and a per protocol at their baseline appointment. Both subjects and field approach. For the latter, compliance was defined as achiev- assistants were blinded. Compliance was assessed by tablet ing an increase in urinary genistein of 100 ng=mmol Cr or counts and urinary assays of genistein and daidzein at urinary daidzein of 200 ng=mmol Cr over baseline. A P-value baseline and 6 weeks using high performance liquid chro- less than 0.05 (two-tailed) or a 95% confidence interval not matography as previously described (Dalais et al, 1998). including the null point were regarded as statistically sig- Weight was measured to the nearest 0.1 kg (with shoes, nificant. All statistical analyses were performed on SPSS socks and bulky clothing removed) using a single pair of version 9.0 for Windows (Cary, NC, USA). electronic scales (Seca Delta Model 707), which were cali- brated using a known weight at the beginning of each clinic. Height was measured to the nearest 0.1 cm (with shoes and Results socks removed) using a Leicester Height Measure (Child A total of 136 boys enrolled in the trial. Four were outside Growth Foundation). One observer (KH) performed all the age limits and deemed ineligible. During the course of physical measurements. Tanner pubertal stage was self- the trial four more subjects withdraw. A total of 128 boys assessed by questionnaire by using drawings made from completed the trial period: 69 in the treatment group and 59 Tanner’s photographs illustrating the five stages of pubertal in the placebo group. In a test of subject knowledge of development. This approach has good agreement with actual allocation at the end of the trial period, 11.6 and 10.2% of examination in males (ICC 0.63; Morris & Udry, 1980). subjects in the treatment and placebo groups, respectively, Bone turnover was assessed at baseline and 6 weeks by two correctly ascertained which group they were assigned to. The assays on first void morning urine samples. Bone formation remainder indicated that they did not know which group

Table 1 Characteristics of participants at baselinea

Study factor Isoflavone (n ¼ 69) Placebo (n ¼ 59) P-value for difference

Age (y) 16.7 (0.78) 16.8 (0.76) 0.68 Weight (kg) 73.4 (13.8) 72.8 (9.7) 0.77 Height (cm) 177.9 (6.8) 177.4 (6.8) 0.67 Bone-specific alkaline phosphatase at baseline (IU) 46.3 (21.2) 49.0 (24.9) 0.50 Pyridinoline creatinine ratio at baseline (nmol=mmol Cr) 8.72 (3.44) 9.29 (3.85) 0.36 Urinary genistein (ng=mmol Cr) 34 (53) 25 (41) 0.27 Urinary daidzein (ng=mmol Cr) 66 (110) 55 (109) 0.58

aData are presented as mean (standard deviation).

European Journal of Clinical Nutrition Phytoestrogens and bone turnover in boys G Jones et al 326 Table 2 Effect of isoflavone supplementation on change over 6 weeks in urinary phytoestrogens, bone turnover markers and growtha

Study factor Isoflavone (n ¼ 69) Placebo (n ¼ 59) P-value for difference

Urinary daidzein (ng=mmol Cr) 466 (565) 31 (118) < 0.001 Urinary genistein (ng=mmol Cr) 250 (322) 16 (63) < 0.001 Bone specific alkaline phosphatase (units) 9.69 (9.04) 8.37 (10.1) 0.47 Pyridinoline creatinine ratio (nmol=mmol creatinine) 0.36 (2.94) 0.77 (2.36) 0.39 Height change (cm) 0.23 (0.81) 0.15 (1.08) 0.65 Weight change (kg) 0.66 (1.49) 0.62 (1.79) 0.91

aAll measures are mean (s.d.).

they were in (76.8 and 83.1%) or they guessed incorrectly tion in children (Marowska et al, 1996), suggesting that (11.6 and 6.7%). There was no significant difference in the they may be a valid short-term marker of bone changes. level of knowledge of the groups (w2 ¼ 0.07, P ¼ 0.80). Demo- Furthermore, animal models have suggested that the graphic and study factors are presented in Table 1. There mechanism by which isoflavones prevent bone loss is by were no significant differences in baseline factors between suppression of bone turnover (Ishida et al, 1998; Tsutsumi, those assigned to isoflavone or placebo. Children were either 1995). They also predict the occurrence of fractures Tanner stage 4 (n ¼ 61) or 5 (n ¼ 67) at baseline. Sixty percent (van Daele et al, 1996) and longer-term changes in bone of children increased their height (by any amount) during density but not accurately enough to be of use in an indivi- the study timeframe. dual patient (Delmas et al, 2000). While the correlation There were significant increases in both the pyridinoline between bone turnover markers and development of bone creatinine ratio (PYR; difference 0.58, 95% CI 0.12 – 1.05) mineralization in children is unknown, our results do not and bone-specific alkaline phosphatase (BAP; difference 8.95 strongly support the need for further bone density units, 95% CI 7.12 – 10.78) during the study timeframe, studies. However, it remains possible that isoflavones have possibly due to the effect of winter. Isoflavone supplementa- an effect on bone mineralization independent of bone turn- tion increased both urinary daidzein and genistein levels over markers. compared to placebo (both P < 0.001), however, there were The phytoestrogen supplement we chose in this trial was no significant differences between the two groups in PYR, clearly effective at increasing urinary excretion of genistein BAP, height change or weight change (Table 2). and daidzein. These agents are relatively weak binders to Per protocol analysis also revealed no significant differ- estrogen receptors (Kuiper et al, 1998), but levels of exposure ences between the two groups in BAP (þ0.53 IU, 95% CI are also higher. It may be that a higher dose would have 73.32, þ4.37) or PYR (þ0.35 nmol=mmol creatinine, 95% CI benefit as in the adult studies (Alekel et al, 2000) or different 70.61, þ1.32). Lastly baseline genistein or daidzein levels in preparations such as coumestrols may have larger effects due urine did not correlate with either BAP or PYR at baseline to their stronger estrogen binding capacity (Kuiper et al, (maximum r 70.07, P ¼ 0.40) and follow-up levels did not 1998). Indeed, one study in rats showed a significant effect correlate with follow up bone turnover markers (maximum r of coumestrol but not isoflavone supplementation on bone 70.11, P ¼ 0.25). loss (Draper et al, 1997). While the use of a valid randomized design is a strength, this study has a number of potential limitations. Firstly, this Discussion study only included adolescent males due to their lower In this randomized trial, supplementation with isoflavones endogenous estrogen levels and the observation that low to the level of usual Japanese intakes markedly increased dose estradiol supplementation can accelerate ulnar growth urinary genistein and daidzein (indicating reasonable com- in boys (Caruso-Nicoletti et al, 1985). It is not possible to pliance), but had no discernible effect on bone formation, generalize to growing females; however, our results would bone resorption or short-term growth in either intention to suggest that supplementation may also be ineffective due treat or per protocol analysis. their greater endogenous estrogen levels and hence greater Assessment of bone health in children is more difficult competition for receptor binding. A time frame of 6 weeks due the combined effects of bone modelling and remodel- may not be sufficient time to observe an effect on bone ling on bone turnover markers and size artefacts on bone turnover; however, studies have shown a rapid response to density. Nevertheless, bone turnover markers do appear to both growth hormone (Lieuw-A-Fa et al, 1995) and estrogen have clinical utility in children (Szulc et al, 2000). They (Kuiper et al, 1998) in children and markers of bone resorp- predict longitudinal growth in infants up to 18 months of tion change in line with daily salt intake ( Jones et al, 1997), age (Lieuw-A-Fa et al, 1995) and children 4 – 18 y (Rauch et al, and within 1 month in adult studies of antiresorptives 1994) as well as response to growth hormone supplementa- (Ettinger et al, 1999; Schnitzer et al, 2000). Alternatively, it

European Journal of Clinical Nutrition Phytoestrogens and bone turnover in boys G Jones et al 327 may be that whole soy protein (containing isoflavones) may Kuiper GG, Lemmen JG, Carlsson B, Corton JC, Safe SH, van der Saag PT, van der Burg B & Gustafsson JA (1998): Interaction of have biologically different actions to isolated isoflavones in a estrogenic chemicals and phytoestrogens with estrogen receptor tablet formulation. Lastly, the sample size may have been beta. Endocrinology 139, 4252 – 4263. insufficient. However, based on our data, we calculate that Lieuw-A-Fa M, Sierra RI & Specker BL (1995): Carboxy-terminal there was 80% power to detect a 16% difference in BAP and a propeptide of human type I collagen and pyridinium cross-links as markers of bone growth in infants 1 to 18 months of age. J. Bone 14% difference in PYR between groups. Smaller changes than Miner. Res. 10, 849 – 853. this are unlikely to be of clinical significance as changes in Marowska J, Kobylinska M, Lukaszkiewicz J, Talajko A, Rymkiewicz- the order of 30 – 50% have been observed in clinical trials of Kluczynska B & Lorenc RS (1996): Pyridinium crosslinks of col- both raloxifene and alendronate in women (Ettinger et al, lagen as a marker of bone resorption rates in children and adolescents: normal values and clinical application. Bone 19, et al 1999; Schnitzer , 2000). 669 – 677. In conclusion, this short-term clinical trial of isoflavone Morris NM & Udry JR (1980): Validation of a self-administered supplementation demonstrates no significant effect on markers instrument to assess stage of adolescent development. J. Youth of bone resorption or short-term growth in adolescent boys. Adoles. 9,271– 280. Picherit C, Chanteranne B, Bennetau-Pelissero C, Davicco MJ, Lebecque P, Barlet JP & Coxam V (2001): Dose-dependent bone- sparing effects of dietary isoflavones in the ovariectomised rat. References Br. J. Nutr. 85, 307 – 316. Alekel DL, Germain AS, Peterson CT, Hanson KB, Stewart JW & Potter SM, Baum JA, Teng H, Stillman RJ, Shay NF & Erdman JW Jr Toda T (2000): Isoflavone-rich soy protein isolate attenuates (1998): Soy protein and isoflavones: their effects on blood lipids bone loss in the lumbar spine of perimenopausal women. Am. J. and bone density in postmenopausal women. Am. J. Clin. Nutr. 68, Clin. Nutr. 72, 844 – 852. 1375S – 1379S. Caruso-Nicoletti M, Cassorla F, Skerda M, Ross JL, Loriaux DL & Rauch F, Schonau E, Woitge H, Remer T & Seibel M (1994): Urinary Cutler GB Jr (1985): Short term, low dose estradiol accelerates excretion of hydroxy-pyridinium cross-links of collagen reflects ulnar growth in boys. J. Clin. Endocrinol. Metab. 61, 896 – 898. skeletal growth velocity in normal children. Exp. Clin. Endocrinol. Dalais FS, Rice GE & Wahlqvist ML (1998): Effects of dietary phyto- 102,94– 97. estrogens in postmenopausal women. Climacteric 1, 124 – 129. Schnitzer T, Bone HG, Crepaldi G, Adami S, MCClung M, Kiel D, Delmas PD, Hardy P, Garnero P & Dain M (2000): Monitoring Felsenberg D, Recker RR, Tonino RP, Roux C, Pinchera A, Foldes AJ, individual response to hormone replacement therapy with bone Greenspan SL, Levine MA, Emkey R, Santora AC Jr, Kaur A, markers. Bone 26, 553 – 560. Thompson DE, Yates J & Orloff JJ (2000): Therapeutic equivalence Draper CR, Edel MJ, Dick IM, Randall AG, Martin GB & Prince RL of alendronate 70 mg once-weekly and alendronate 10 mg daily in (1997): Phytoestrogens reduce bone loss and bone resorption in the treatment of osteoporosis. Aging (Milano) 12,1– 12. oophorectomized rats. J. Nutr. 127, 1795 – 1799. Somekawa Y, Chiguchi M, Ishibashi T & Aso T (2001): Soy intake Ettinger B, Black DM, Mitlak BH, Knickerbocker RK, Nickelsen T, related to menopausal symptoms, serum lipids, and bone mineral Genant HK, Christiansen C, Delmas PD, Zanchetta JR, Stakkestad J, density in postmenopausal Japanese women. Obstet. Gynecol. 97, Gluer CC, Krueger K, Cohen FJ, Eckert S, Ensrud KE, Avioli LV, Lips 109 – 115. P & Cummings SR (1999): Reduction of vertebral fracture risk in Szulc P, Seeman E & Delmas PD (2000): Biochemical measurements postmenopausal women with osteoporosis treated with raloxi- of bone turnover in children and adolescents. Osteopor. Int. 11, fene: results from a 3-year randomized clinical trial. Multiple 281 – 294. Outcomes of Raloxifene Evaluation (MORE) Investigators. JAMA Tsutsumi N (1995): Effect of coumestrol on bone metabolism in 282, 637 – 645. organ culture. Biol. Pharm. Bull. 18, 1012 – 1015. Falahati-Nini A, Riggs BL, Atkinson EJ, O’Fallon WM, Eastell R & Upmalis DH, Lobo R, Bradley L, Warren M, Cone FL & Lamia CA Khosla S (2000): Relative contributions of testosterone and estro- (2000): Vasomotor symptom relief by soy isoflavone extract tablets gen in regulating bone resorption and formation in normal elderly in postmenopausal women: a multicenter, double-blind, rando- men. J. Clin. Invest. 106, 1553 – 1560. mized, placebo-controlled study. Menopause 7,236– 242. Fanti P, Monier-Faugere MC, Geng Z, Schmidt J, Morris PE, Cohen van Dale PL, Seibel MJ, Burger H, Hofman A, Grobbee DE, van D & Malluche HH (1998): The phytoestrogen genistein reduces Leeuwen .JP, Birkenhager JC & Pols HA (1996): Case control bone loss in short-term ovariectomized rats. Osteoporos Int. 8, analysis of bone resorption markers, disability and hip fracture 274 – 281. risk: the Rotterdam study. Br. Med. J. 312, 482 – 483. Ishida H, Uesugi T, Hirai K, Toda T, Nukaya H, Yokotsuka K & Tsuji K Wang MC, Aguirre M, Bhudhikanok GS, Kendall CG, Kirsch S, (1998): Preventive effects of the plant isoflavones, daidzein and Marcus R & Bachrach LK (1997): Bone mass and hip axis length genistein, on bone loss in ovariectomized rats fed a calcium- in healthy Asian, black, Hispanic, and white American youths. deficient diet. Biol. Pharm. Bull. 21,62– 66. J. Bone Miner. Res. 12, 1922 – 1935. Jones G, Beard T, Parameswaran V, Greenaway T & Von Witt R (1997): Wangen KE, Duncan AM, Merz-Demlow BE, Xu X, Marcus R, A population-based study of the relationship between salt intake, Phipps WR & Kurzer MS (2000): Effects of soy isoflavones on bone resorption and bone mass. Eur. J. Clin. Nutr. 51, 561 – 565. markers of bone turnover in premenopausal and postmenopausal Kimira M, Arai Y, Shimoi K & Watanabe S (1998): Japanese intake of women. J. Clin. Endocrinol. Metab. 85, 3043 – 3048. flavonoids and isoflavonoids from foods. J. Epidemiol. 8, 168 – 175.

European Journal of Clinical Nutrition