Effects of Genistein and Hormone-Replacement Therapy on Bone Loss in Early Postmenopausal Women: A Randomized Double-Blind Placebo-Controlled Study


  • The authors have no conflict of interest.


The natural isoflavone phytoestrogen genistein has been shown to stimulate osteoblastic bone formation, inhibit osteoclastic bone resorption, and prevent bone loss in ovariectomized rats. However, no controlled clinical trial has been performed so far to evaluate the effects of the phytoestrogen on bone loss in postmenopausal women. We performed a randomized double-blind placebo-controlled study to evaluate and compare with hormone-replacement therapy (HRT) the effect of the phytoestrogen genistein on bone metabolism and bone mineral density (BMD) in postmenopausal women. Participants were 90 healthy ambulatory women who were 47–57 years of age, with a BMD at the femoral neck of <0.795 g/cm2. After a 4-week stabilization on a standard fat-reduced diet, participants of the study were randomly assigned to receive continuous HRT for 1 year (n = 30; 1 mg of 17β-estradiol [E2] combined with 0.5 mg of norethisterone acetate), the phytoestrogen genistein (n = 30; 54 mg/day), or placebo (n = 30). Urinary excretion of pyridinoline (PYR) and deoxypyridinoline (DPYR) was not significantly modified by placebo administration either at 6 months or at 12 months. Genistein treatment significantly reduced the excretion of pyridinium cross-links at 6 months (PYR = −54 ± 10%; DPYR = −55 ± 13%; p < 0.001) and 12 months (PYR = −42 ± 12%; DPYR = −44 ± 16%; p < 0.001). A similar and not statistically different decrease in excretion of pyridinium cross-links was also observed in the postmenopausal women randomized to receive HRT. Placebo administration did not change the serum levels of the bone-specific ALP (B-ALP) and osteocalcin (bone Gla protein [BGP]). In contrast, administration of genistein markedly increased serum B-ALP and BGP either at 6 months (B-ALP = 23 ± 4%; BGP = 29 ± 11%; p < 0.005) or at 12 months (B-ALP = 25 ± 7%; BGP = 37 ± 16%; p < 0.05). Postmenopausal women treated with HRT had, in contrast, decreased serum B-ALP and BGP levels either at 6 months (B-ALP = −17 ± 6%; BGP = −20 ± 9%; p < 0.001) or 12 months (B-ALP = −20 ± 5%; BGP = −22 ± 10%; p < 0.001). Furthermore, at the end of the experimental period, genistein and HRT significantly increased BMD in the femur (femoral neck: genistein = 3.6 ± 3%, HRT = 2.4 ± 2%, placebo = −0.65 ± 0.1%, and p < 0.001) and lumbar spine (genistein = 3 ± 2%, HRT = 3.8 ± 2.7%, placebo = −1.6 ± 0.3%, and p < 0.001). This study confirms the genistein-positive effects on bone loss already observed in the experimental models of osteoporosis and indicates that the phytoestrogen reduces bone resorption and increases bone formation in postmenopausal women.


Osteoporosis is a major public health problem in the elderly population around the world, particularly in women. The risk of developing osteoporosis after menopause is determined largely by the peak bone mass obtained at young adult age and by the rate of bone loss after menopause. Postmenopausal bone loss is caused primarily by the sharp decrease in estrogen levels that is inherent to menopause.(1)

Hormone-replacement therapy (HRT) is effective in the prevention of bone loss in early menopause, but it also is accompanied by several adverse effects including uterine bleeding and hyperplasia.(2) Indeed, estrogens are combined with low-dose progestins to prevent estrogen-induced endometrial cancer.(3) However, there is evidence that the use of estrogen, with or without progestin, increases the risk of breast cancer.(4) This potential drawback, in combination with the low compliance to HRT, has stimulated in recent years the research of alternatives to the classical HRT.

Phytoestrogenic molecules have received a great deal of attention over the last few years because of their potentially preventive roles against a few of today's most prevalent chronic diseases, namely, osteoporosis, cardiovascular disease, and hormone-related cancers.(5,6) The particular plant molecules of interest are isoflavones, which are found in abundance in soybeans and their derivative foods. Soy protein and soy-derived phytoestrogens prevent bone loss in rat models of postmenopausal osteoporosis.(7–9) Furthermore, it has been suggested that high soy protein intake in postmenopausal women is associated with a higher bone mineral density (BMD).(10)

Of the several isoflavones that are made by soybeans, genistein has been experimentally shown to be the most efficacious, in particular, in animal models.(11) Genistein may interact with nuclear estrogen receptors (ERs) by either activating or inhibiting transcription of cell-specific genes. Cells may vary in their distribution of the two described types of ERs (ER-α and ER-β) depending on tissue. Reproductive cells, especially those of the breast and uterus, are rich in ER-α, whereas other cells such as in bone have greater amounts of ER-β than ER-α. Of particular interest is the relative affinity of ER-β for the phytoestrogen genistein, which is approximately seven times that of ER-α.(12) By binding to ER-β, genistein may have differential effects that may depend on the tissue distribution of ERs; theoretically, genistein might have agonist and positive effects on the bone cells without any significant adverse effect on the breast and uterus. Indeed, experimental evidence confirmed this hypothesis; in fact, genistein administration in ovariectomized mice (an experimental model of postmenopausal osteoporosis) restored the increased B-lymphopoiesis and bone loss caused by estrogen deficiency without exhibiting a substantial effect on the uterus.(13) These results, if clinically confirmed in postmenopausal women, might predict a positive impact in osteoporosis progression and bone fracture event rates. Therefore, the aim of our study was to investigate and compare the effects of genistein and HRT on BMD and metabolism and on the reproductive system of early postmenopausal women.



After the Ethical Committee approved our study, participants were recruited among those reporting to the Center for Osteoporosis in the Department of Internal Medicine of the University of Messina (Messina, Italy). This was a single center, double-blind, placebo-controlled, randomized study. All participants gave informed consent. Participants were healthy, ambulatory women who were 47–57 years of age, had not undergone surgically induced menopause, had not had a menstrual period in the preceding year, and had a follicle-stimulating hormone (FSH) level >50 IU/liter and a serum 17β-estradiol (E2) level of ≤100 pmol/liter.

At the beginning of the study, a complete family history, physical examination, laboratory evaluation (chemistry and hematology panel), and measurements of BMD at the lumbar spine and femoral neck were done at the Department of Internal Medicine of the University of Messina. All the women underwent a mammography and a transvaginal ultrasound study for determination of endometrial thickness

Exclusion criteria were clinical or laboratory abnormalities that suggested cardiovascular, hepatic, or renal disorders; coagulopathy, use of oral or transdermal estrogen, progestin, androgen, or other steroids in the preceding year; smoking habit of more than two cigarettes per day; previous treatment with any drug that could affect the skeleton; a family history of estrogen-dependent cancer; BMD at the femoral neck >0.795 g/cm2. This BMD value corresponds to a T score = −1 SD.


All patients received dietary instruction in an isocaloric fat-restricted diet offering 30% energy from fat, <10% of energy from saturated fatty acids, and a cholesterol intake of <300 mg/day. To avoid any interference with the possible effects of the different therapeutic interventions on lipid profile, the intake of soy products, legumes, or other nutritional supplements was prohibited; moreover, the previous isoflavone intake before randomization, assessed by a food-frequency questionnaire, was 1–2 mg/day. This intake has been shown to be typical of the Western population.(14) Dietary calcium intake was estimated at baseline and after 2, 4, 6, 8, and 10 months by means of a food-frequency questionnaire. Women with a calcium intake of <500 mg/day were advised to increase their intake. This diet was continued throughout the study and compliance was reinforced by a nutritionist.


After a 4-week stabilization on the standard fat-reduced diet, participants to the study were randomly assigned to receive continuous HRT (n = 30; 1 mg/day of E2 combined with norethisterone acetate; Activelle, Norvo Nordisk, Logenhagen, Denmark), the phytoestrogen genistein (n = 30; 54 mg/day), or placebo (n = 30). This amount of genistein chosen in this study is slightly greater than a 50-mg/day intake. Previously, this intake of phytoestrogen was described to have biological effects.(15) Genistein was obtained from Lab. Plants (Messina, Italy). Tablets contained 54 mg of total isoflavone. The purity of genistein was ∼98%. Placebo, HRT, and genistein tablets appeared exteriorly similar.


The BMD of the anteroposterior lumbar spine and femoral neck was measured by DXA (Hologic QDR 4500 W, Techonologic Srl, Torino, Italy) at baseline and after 1 year of treatment. The instrument was calibrated on a daily basis according to the manufacturer's instructions. Reproducibility was calculated as a CV obtained by weekly measurements of a standard phantom on the instrument and by repeated measurements obtained in three patients of different ages. The CV of our instrument is 0.5% with the standard phantom; in vivo we calculated a CV of 1.1% for the lumbar spine, 1.5% for the neck, and 3.2% for Ward's triangle. BMD data were expressed as grams per squared centimeter.

After an overnight fast, venous blood samples were drawn through a polyethylene catheter inserted in a forearm vein between 8 and 9 a.m. The serum was separated from the blood corpuscles by centrifugation and kept frozen at −70°C until analyzed for calcium (Ca2+), bone-specific ALP (B-ALP), osteocalcin (bone Gla protein [BGP]), intact parathyroid hormone (PTH), 25-hydroxyvitamin D3 [25(OH)D3], E2, and FSH. Total cholesterol, high-density lipoprotein cholesterol, low-density lipoprotein cholesterol, and triglyceride levels were measured in the Azienda Policlinico Universitario Immunochemistry Laboratory.

A 2-h fasting morning urine was collected at the same time for measurements of pyridinium cross-links (pyridinoline [PYR] and deoxypyridinoline [DPYR]) and creatinine. These parameters were evaluated at baseline and 6 months and 12 months after treatment.

Ca2+ (normal range, 2.1–2.6 mM) and creatinine (0.13–0.22 mmol kg−1 of body weight/24 h in urine) were determined by automated routine procedures. BGP (normal range, 1.6–17.4 ng/ml), B-ALP (normal range, 8.5–17.9 μg/liter), PTH (normal range, 1.2–7.2 pmol/liter), 25(OH)D3 (normal range, 25–125 nM), and FSH (normal range, 21–153 IU/liter in postmenopausal phase) were measured by an immunoradiometric method (BOUTY SpA; Italiana Laboratori Bouty, Milan, Italy). E2 (normal range, 37–110 pmol/liter in postmenopausal phase) were evaluated using a solid-phase immunoassay (Roche Diagnostics, Milan, Italy). PYR (normal range, 25–91 pmol/μmol of urinary creatinine) and D-PYR (normal range, 3–21 pmol/μmol of urinary creatinine) were measured by an HPLC (Bio-Rad Laboratories Diagnostics, Richmond, CA, USA).

To measure genistein plasma levels, blood samples (0.5 ml) were collected in polypropylene tubes containing 50 μl of heparin (50,000 IU) and after centrifugation at 3000g at 4°C for 10 minutes, each sample was stored at −70°C until analysis. The assay was performed by using an HPLC method with UV detection(16) with some modifications. Briefly, 0.5 ml of plasma was added in polycarbonate tubes containing 4-hydroxybenzophenone, used as an internal standard, and terbutylmethyl ether. After shaking, the organic layer was recovered and evaporated in a vacuum concentrator system (Heto La. Equipment, Allerød, Denmark). The extract was reconstituted with methanol 0.05 M ammonium acetate buffer, pH 4. 5 (35:65, vol/vol), and 50 μl of the solution was injected into the HPLC apparatus. The column used was a 3 μm Luna C8, 150 × 4.6 mm internal diameter (id; Phenomenex, Terrance, CA, USA). Chromatography was carried out at room temperature using a mobile phase of acetonitrile-0.05 M ammonium formate buffer, pH 4.0 (28:75, vol/vol), at a flow rate of 1.0 ml/minute. The HPLC equipment consisted of a solvent delivery module (Mod 422 Master; Kontron Instruments, Everett, WA, USA), a programmable variable wavelength detector (Spectromonitor 4100; Thermo Separation Products, Riviera Beach, FL, USA), connected to an automatic integrator (Mod. CR-3A; Shimadzu, Kyoto, Japan). The UV detector was set at a wavelength of 260 nm. The concentration of plasma genistein was expressed in micromoles per liter.

Clinic visits

The women were questioned about any symptoms at clinic visits every 3 months. Standard clinical evaluations and laboratory analyses, including hematological, renal, and liver function tests, were performed every 6 months. Transvaginal uterine ultrasound and PAP tests were performed at baseline and after 6 months and 12 months, whereas mammography was performed at baseline and after 1 year. All unfavorable and unintended clinical effects were considered adverse effects and were evaluated by the investigators with respect to severity, duration, seriousness, and relation to the study drug and outcome.


The primary evaluation of the efficacy data according to the intention to treat included all 90 postmenopausal women in whom BMD was measured at baseline and after 1 year of treatment.

The effect of treatment on BMD was assessed by ANOVA. The evaluation of the incidence of side effects in the several groups of postmenopausal women was carried out with the Fisher's exact probability test. All statistical tests were two sided. All data are reported as means and SD. A value of p < 0.05 was considered statistically significant. Stepwise linear regression was used to select the independent predictors of increased BMD for the final best multivariate models. Variables were included in the final model if the value of p < 0.05. Statistical analysis was performed using SPSS, Inc. for Windows release 6.0 (SPSS Inc., Chicago, IL, USA).


The baseline characteristics of the postmenopausal women participating in our study are shown in Table 1. No statistically significant difference was observed among the different groups. The mean femoral neck BMD value in the three randomized groups was lower than 0.795 g/cm2, suggesting the presence of significant bone loss. The serum levels of E2, FSH, calcium, PTH, 25(OH)D3, and the plasma levels of genistein as well as the markers of bone formation (B-ALP and BGP) and resorption (PYR and DPYR) were similar among the three different groups. Total and low-density lipoprotein cholesterol levels, high-density lipoprotein cholesterol levels, and triglycerides also were similar among the three different groups (results not shown).

Table Table 1.. Baseline Characteristics of 90 Participants in the Trial (Mean ± SD)
original image

Treatment with placebo, genistein, or HRT did not significantly modify the serum levels in calcium, PTH, and 25(OH)D3 either after 6 months or 12 months of therapy (Table 2). With genistein or placebo, total and low-density lipoprotein cholesterol levels, high-density lipoprotein cholesterol levels, and triglycerides did not change. HRT significantly reduced total cholesterol and low-density lipoprotein (p < 0.05) and it also significantly increased high-density lipoprotein (p < 0.05). No effect of HRT was observed in triglyceride levels (results not shown).

Table Table 2.. Biochemical Parameters in Postmenopausal Women
original image

Administration of 54 mg/day of genistein resulted in a marked increase in plasma levels of the phytoestrogen, whereas the phytoestrogen did not change the plasma levels of E2 (Table 2). HRT significantly increased the circulating levels of E2 while it did not affect the plasma levels of the phytoestrogen genistein (Table 2). Placebo treatment did not modify plasma E2 or the circulating levels of genistein (Table 2).

Urinary excretion of PYR and DPYR was not significantly modified by 6 months and 12 months of placebo administration (Table 2). The daily administration of genistein significantly reduced the urinary excretion of pyridinium cross-links at 6 months (PYR = −54 ± 10%; DPYR = −55 ± 13%; p < 0.001) and 12 months (PYR = −42 ± 12%; DPYR = −44 ± 16%; p < 0.001) (Table 2). A similar and not statistically different (compared with the genistein-treated group) decrease in the excretion of PYR and DPYR was observed in the postmenopausal women randomized to receive HRT (Table 2).

Placebo administration did not change the serum levels of B-ALP. In contrast, the daily administration of 54 mg of the phytoestrogen genistein caused a marked increase in the serum levels of serum B-ALP and BGP either at 6 months (B-ALP = 23 ± 4%; BGP = 29 ± 11%; p < 0.005) or 12 months (B-ALP = 25 ± 7%; BGP = 37 ± 16%; p < 0.05) (Table 2). Postmenopausal women treated with HRT showed, in contrast, a marked reduction in the serum levels of B-ALP and BGP levels either at 6 months (B-ALP = −17 ± 6%; BGP = −20 ± 9%; p < 0.001) or 12 months (B-ALP = −20 ± 5%; BGP = −22 ± 10%; p < 0.001) months (Table 2).

BMD studies performed after 1 year of treatment with placebo, genistein, and HRT confirmed the biochemical results. Indeed, at the end of experimental period, genistein and HRT significantly increased BMD in the femoral neck (genistein = 3.6 ± 3%; HRT = 2.4 ± 2%; placebo = −0.65 ± 0.1%; p < 0.001), in Ward's triangle (genistein = 4 ± 2.7%; HRT = 3 ±2; placebo = −0.36 ± 0.5%; p < 0.0006), and in the lumbar spine (genistein = 3 ± 2%; HRT = 3.8 ± 2.7%; placebo = −1.6 ± 0.3%; p < 0.04) when compared with placebo (Fig. 1).

Figure FIG. 1.

Mean (±SEM) changes from baseline in BMD of the lumbar spine, femoral neck, and Ward's triangle (#p < 0.01 versus placebo).

In a multivariate model, the independent predictors of increased BMD were initial BMD (β = 1.7; SE = 0.17; R2 = 0.92; p < 0.001) and urinary DPYR levels (β = −0.12; SE = 0.001; R2 = 0.92; p < 0,05) for lumbar spine; initial BMD (β = 0.64; SE = 0.12; R2 = 0.88; p < 0,001) and DPYR (β = −0.38; SE = 9.2; R2.= 0.88; p < 0.001) for Ward's triangle; initial BMD (β = 0.91; SE = 0.02; R2.= 0.97; p < 0.001), urinary DPYR levels (β = −0.23; SE = 2; R2 = 0.97; p < 0.001) and serum B-ALP (β = 0.43; SE = 2.5; R2.= 0.97; p < 0.001) for femoral neck (Table 3). Similar results were observed in HRT-treated women, except for the data regarding markers of bone formation; in fact, these parameters were negatively correlated with final BMD (Table 3).

Table Table 3.. Independent Predictors of Final BMD in Multivariate Analysis Among Postmenopausal Women Treated With Genistein and HRT for 1 Year
original image

Genistein, HRT, and placebo were generally well tolerated and ingested with a high degree of compliance. There were no significant changes in routine biochemistry, liver function, or hematology results. Side effects and menopausal complaints in postmenopausal women randomized to receive placebo, genistein, or HRT are summarized in Table 4. The daily administration of 54 mg of the phytoestrogen genistein did not cause any significant change in the endometrial thickness and in the breast (no patient in any of the groups showed significant change in the mammography exams at 1 year of follow-up). Vaginal bleeding was present only in a significant number of patients taking HRT. Breast tenderness was significantly present in the women randomized to receive HRT, whereas hot flushes were markedly reduced in women given HRT (Table 4).

Table Table 4.. Summary of Side Effects and Menopausal Complaints in Postmenopausal Women Randomized to Receive Placebo Genistein or HRT
original image


This study clearly shows that genistein prevents bone loss caused by estrogen deficiency without exerting substantial adverse effects on the uterus and breast. In addition, it reduces bone resorption markers and enhances new bone formation parameters, in turn producing a net gain of bone mass.

The effect of genistein on bone resorption was determined by the measurement of urinary excretion of PYR and DPYR. In our patients, urinary levels of these cross-links, derived from lysine and hydroxylysine residues (which are produced only during the formation of mature collagen(17)) were significantly reduced by administration of genistein; thus, the reduction of bone loss is caused by, at least in part, a decrease in bone resorption and this effect is similar to that caused by HRT involving, therefore, ERs.

However, other effects also may explain the genistein effect on bone resorption. In fact, previous studies in cell culture systems have shown that genistein has a direct suppressive effect on osteoclasts and also can induce their apoptosis through the pathway of intracellular Ca2+ signaling.(18)

Moreover, it has recently been shown that genistein causes a significant decrease in the number of osteoclasts in rat femoral tissue,(19) thus suggesting that the suppressive effect of the phytoestrogen on rat bone osteoclasts may involve, at least in part, the inhibition of protein kinase and the activation of protein tyrosine phosphatase. In fact, protein tyrosine kinase inhibition in rat osteoclast elevates cytosolic Ca2+ concentration via activation of a dihydropyridine-insensitive, nonspecific Ca2+ entry pathway, in turn disrupting the formation of actin rings.(20)

Furthermore, genistein has been found to possess a strong inhibitory effect on osteoclast-like cell formation in mouse marrow culture, and this inhibitory action may also involve cyclic adenosine monophosphate (cAMP) signaling.(21)

All these genistein-mediated effects might explain the inhibition of bone resorption observed in our postmenopausal women treated with the phytoestrogen.

Interestingly, in our patients treated with genistein, we also observed a significant enhancement in B-ALP and osteocalcin, both markers of osteoblast activity. This genistein effect on bone formation strongly differs from the well-known HRT effect on this tissue; in fact, women assigned to HRT treatment showed a marked reduction in markers of bone formation.

This result, produced for the first time in postmenopausal women, is in close agreement with a number of previously published preclinical studies that indicate that genistein has an anabolic effect on bone metabolism in rats.(21–26) More specifically, in ovariectomized rats the administration of genistein was associated with a higher bone formation rate per tissue volume and with a significant trend toward a higher number of osteoblasts per bone perimeter.(27)

The mechanism by which genistein affects B-ALP activity is largely unknown. This effect is completely abolished by the presence of cycloheximide, an inhibitor of protein synthesis, in a tissue culture system in vitro,(28) suggesting that this genistein effect is based on a “de novo” protein synthesis. The stimulatory effect on protein synthesis has also been shown by the activation of aminoacyl-tRNA synthethase in osteoblastic MC3T3-E1 cells.(29) Yamaguchi and coworkers(28) speculated that the phytoestrogen effect on B-ALP might be caused in part by either the stimulation of the ER protein synthesis or the amplification of ER complex interaction with nuclear DNA in osteoblastic cells. In fact, these cells express both ER-α and ER-β(30) and it is likely that genistein may act on bone by a mechanism involving ER-β, because the phytoestrogen has been reported to bind ER-β in bone cells with a higher affinity and lower capacity than estrogen.(31) Collectively, these experimental evidences suggest that the phytoestrogen effect on bone formation may be a consequence, at least in part, of a genomic and ER-mediated effect.

However, the effects of genistein on bone formation need not be restricted to a direct hormonal action; in fact, genistein also has been shown to have a strong inhibitory effect on protein tyrosine kinases(32,33) in human bone cells. Moreover, the phytoestrogen significantly stimulates thymidine incorporation and increases cell number in human vertebrae-derived bone cells,(34) an effect that is not related to its estrogenic properties but to its inhibitory action on soluble tyrosine kinases.

Indeed, there is no evidence that estrogens stimulate human bone cell proliferation; in fact, estrogen treatment clearly has been shown to inhibit both basal and growth factor-stimulated proliferation of normal rodent and human bone cells as well as human osteosarcoma cells in vitro. (35,36) In contrast, low doses (5–20 μM) of genistein not only did not inhibit, but significantly stimulated basal and epidermal growth factor (EGF)-induced human bone cell proliferation.(37)

All these findings are intriguing because these observations are in conflict with the concept that activation of protein tyrosine kinase is required for cell proliferation.(37) However, genistein has been shown to inhibit both basal- and growth factor-induced proliferation of several normal and cancer cell lines(38,39); therefore, it is possible that the mitogenic effect of genistein may be unique to bone cells.

These composite effects of genistein on bone turnover produced a net gain in BMD in our postmenopausal women. As a matter of fact, we observed a significant increase in lumbar spine and femoral neck BMD in the genistein group. This improvement was similar to that observed in the HRT group at the lumbar spine level, but it was slightly greater than that caused by HRT at the femoral neck.

In a multivariate analysis the independent predictors on final femoral neck BMD in postmenopausal women treated with genistein were in addition to greater baseline BMD, elevated serum B-ALP levels, and decreased urinary DPYR concentration. In contrast, the change in BMD of the HRT group was linked to a decrease in all markers of bone turnover.

Genistein treatment for 1 year was not associated with any significant adverse effect on the breast and uterus. This confirms the good safety profile of genistein in postmenopausal women. This finding, together with the positive effect of genistein on bone, might lead to us recommend the use of this phytoestrogen for the prevention of bone loss in premenopausal women, without significant adverse effects. In fact, the increased phytoestrogen intake in premenopausal women is also useful for preventing several hormone-related cancer disease through the suppression of gonadotrophin output, which leads to a lengthening of the menstrual cycle and, consequently, a lower lifetime exposure to E2 levels.(40) Furthermore, increased phytoestrogen intake causes a decrease in urinary estrogen excretion likely because of a reduced estrogen synthesis and a reduced production of genotoxic estrogen metabolites.(41)

In conclusion, this study confirms for the first time in postmenopausal women the important role of genistein in the prevention of bone loss induced by estrogen deficiency.