Epidemiological data suggest that populations with high intakes of soy, i.e. Asians, have a lower incidence of osteoporotic fractures [16,17]. Asian women typically consume about 20 g of soy daily which provides approximately 40 mg isoflavones [18,19]. However, lower rates of fractures in these populations may not be fully attributed to soy consumption as there are a number of other confounding factors which can influence skeletal health.

From the research point of view, there are a number of animal studies which have shown that soy protein and/or its isoflavones positively influence bone mineral density (BMD) [8,20-26]. In terms of human studies, there are limited numbers of trials that have examined the effects of soy and its isoflavones on bone. Some of these clinical trials [27-29] are of short duration varying between 3 to 6 months, making the findings questionable since periods less than one year may not be sufficient to detect clinically relevant changes in bone mass. Nonetheless, even the findings of the few clinical studies of one to two year duration that have been conducted [30-34] are inconclusive. For instance, Vitolins et al. [30] reported that daily consumption of 25 g soy protein with 5, 42 or 58 mg isoflavones had no bone preserving effects in peri- and post-menopausal women in a two-year study. Recent findings from two groups [31,32] have shown similar effects of soy protein and its isoflavones on bone. In contrast, other studies have suggested that isoflavone-rich soy milk delivering 80 to 90 mg isoflavones [33] or soy products delivering 40 to 60 mg isoflavones on a daily basis [34] have some bone protective effects.

In the present study, the daily consumption of 25 g protein for one year irrespective of the source resulted in no significant changes in hip BMD and BMC. Other investigators [51,52] have reported that diets high in protein were associated with higher BMD in femoral neck. We speculate that higher dietary protein may have a protective effect on hip BMD over the long term. This notion, however, seems somewhat paradoxical because high protein diets, especially proteins rich in sulfur-containing amino acids, are known to increase urinary calcium that may result in accelerated bone loss [53]. Nonetheless, a counter-argument has been made that protein-associated hypercalciuria is due to enhanced intestinal calcium absorption and not the breakdown of bone [54,55].

Our findings do not support a bone protective role for soy protein and its isoflavones at the level used in this study. Whether higher amount of soy protein and/or its isoflavones can reverse bone loss remains to be illustrated. Nonetheless, higher doses of soy protein with varying levels of isoflavones have not consistently shown to exert beneficial effects on bone. For instance, Gallagher et al. [32] supplemented the diets of postmenopausal women for nine months with 40 g soy protein delivering three levels of isoflavones (0, 52, and 96 mg) but all three groups experienced bone loss. On the other hand, six-month studies by Potter et al. [28] and Alekel and colleagues [29] reported positive effects of soy protein supplementation on BMD. Potter et al. [28] showed that 40 g of soy protein containing 90 mg isoflavones was able to attenuate lumbar spine (L1-L4) BMD, however, the same amount of protein with 56 mg isoflavones had no such an effect. Although Alekel et al. [29] suggested that 40 g soy protein supplementation with 80 mg isoflavones was able to attenuate bone loss from lumbar spine, women still lost 0.2% BMD in six months. Their data [29] imply that soy protein or its isoflavones are incapable of increasing bone mass in perimenopausal women.

As for an effect of soy isoflavones alone, Chen et al. [56] recently reported that supplementing postmenopausal women with soy isoflavones (40 and 80 mg/d) for one year resulted in favorable increases in BMC of the hip in women who are at least four years postmenopausal and are of low body weight or have low levels of dietary calcium. Similar to that study [56], the majority of the women in our present study were four or more years postmenopausal; however, our study participants had adequate calcium intakes and did not have low body weights. It is possible that this difference in the nutrition status of the study participants between the two studies may be responsible for the discrepancy in the observed effects on bone.

As far as which of the many isoflavones in soy is responsible for the effects on bone, to date, the most convincing data on the effect of a single isoflavone, genistein, on bone have been reported in a one-year study by Morabito and colleagues [57]. They demonstrated that both genistein at a dose of 54 mg/d and HRT increased BMD in early postmenopausal women. In that study [57], genistein significantly increased BMD of the femoral neck by 3.6% and lumbar spine by 3.0% while HRT increased femoral neck and lumbar spine BMD by 2.4 and 3.8%, respectively. These authors [57] suggested that genistein reduces bone resorption markers and enhances new bone formation parameters resulting in a net gain of bone mass. Isolated isoflavones derived from other sources such as red clover have also been found to positively affect bone. For example, Clifton-Bligh and colleagues [58] reported that clover-derived isoflavones at doses of 57 and 85 mg isoflavones/day were able to significantly increase BMD of the proximal radius and ulna by 4.1 and 3%, respectively after 6 months. Similarly, Atkinson et al. [59] showed that red-clover derived isoflavones (26 mg biochanin A, 16 mg formononetin, 1 mg genistein, and 0.5 mg daidzein) slowed the loss of lumbar spine BMC and BMD. These data suggest that isoflavones from sources other than soy, also have osteoprotective effects.

In the present study, biomarkers of bone formation, i.e. osteocalcein, IGF-I, and BSAP were all significantly elevated in both groups. However, the specific marker of bone resorption, urinary Dpd, was not altered. A number of clinical studies have evaluated the effects of soy protein with its isoflavones on bone biomarkers. Overall, the effects of soy and/or its isoflavones have produced no consistent effects on biomarkers of bone turnover. For instance, biomarkers of bone formation have been reported to either increase [60,61] or not change [29,62] as result of soy supplementation. Similarly, biomarkers of bone resorption have been reported to decrease [62-64], not change [61] or even increase [60]. We have previously reported that 40 g of soy protein providing 90 mg of isoflavones/day reduced Dpd in postmenopausal women not on HRT [62]. In the present study, participants were asked to consume a lower amount of soy protein (only 25 g providing 60 mg of isoflavones/day). It is possible that the reduced dose of soy in this study may have contributed to the lack of effect on bone resorption (as assessed by Dpd excretion) and that a dose-response study may be justified in order to achieve both increases in markers of bone formation and reductions in markers of bone resorption. Nevertheless, the positive changes in biomarkers of bone formation in the present study have not translated to increases in BMD and BMC. Whether the positive effects of protein supplementation on bone biomarkers would translate to better bones needs to be assessed in a longer term study.

Although in this study, soy supplementation for one year did not produce any estrogenic effects as assessed by circulating sex hormone levels, it did decrease SHBG concentrations Decreases in SHBG result in increases in the availability of circulating estrogens [65]. Thus, soy supplementation may have increased the availability of estrogens without affecting actual concentrations. However, this is speculative and measurement of bioavailable estradiol is necessary to confirm this statement.