Bone Effects
Key Recommendations
* Efficacy studies of soy isoflavones on bone density should be performed for a minimum of 24 months.
* More human studies on bone density need to be conducted with equol producers, soy products with a higher genistein content, and higher doses of isoflavones.
Mechanisms Active on Bone
Genistein may activate ER-β in osteoblasts. Because it is also a tyrosine kinase inhibitor, it inhibits cell growth and acid transport in bone cells leading to inhibition of osteoclast activity. Therefore, soy isoflavones may suppress bone resorption and minimize bone loss under some conditions.
Clinical Outcomes in Animals
The ovariectomized rat model is classically used for studying osteoporosis. Combination isoflavones have generally provided little benefit to bone in ovariectomized rats. In one study, 6-month-old ovariectomized and sham-operated rats were studied with an isolated combination of isoflavones at levels of 0.3 and 0.8 mg/g of diet compared to estrogen. Estrogen, but not soy isoflavones, was found to prevent trabecular bone loss in these rats, and no advantage was found in combining estrogen and soy isoflavones. The study also demonstrated that estrogen, but not soy isoflavones, suppressed bone remodeling as determined by dynamic histomorphometry and kinetic modeling. Calcium and bone balance (bone formation minus bone resorption) were unaffected by soy isoflavones. Another 3-year longitudinal study in 181 ovariectomized female monkeys compared the effect of a diet with or without soy isoflavones to estrogen treatment on spine and whole body bone mineral content and bone mineral density (BMD); the results showed that estrogen, but not soy isoflavones, was found to protect against menopausal bone loss.
Equol was found to be as effective as estradiol in preventing bone loss but without stimulating the uterus. Genistein has also had similar effects in preventing bone loss in several other studies. One of these studies suggested that genistein increased bone formation. The amounts of soy isoflavones used in these animal studies, however, were severalfold greater than those used in human studies.
Clinical Outcomes in Humans
Studies of Asian populations who regularly consume much larger amounts of soy than Caucasians have suggested that a high-soy diet is related to a much lower incidence of osteoporotic fracture. For instance, genistein intake was twofold higher in Japanese than Chinese women and 2,000 times higher than in Caucasian women.
In the US SWAN study, genistein intake was correlated with bone density in premenopausal Japanese-American women. In the top tertile of genistein intake, the adjusted spine bone density was 7.7% higher and femoral neck bone density was 12% higher compared to the lowest tertile of intake.
There have now been 23 human studies on the effect of soy isoflavones on bone density. In general, a decrease in bone density at 12 months is probably a valid indicator of no effect. A positive result within 12 to 18 months, however, may represent the transient remodeling effect of an antiresorptive agent. Therefore, efficacy studies should be performed for a minimum of 24 months. Only four of the 23 studies were properly designed to address the potential efficacy of soy isoflavones to prevent bone loss.
* In the first study, there was a very significant increase in bone density at all bone sites of approximately 5% compared to a 5% loss on placebo in a study comparing pure 54 mg genistein/day to placebo. The results from this study are completely at variance with the above studies that used comparable doses of genistein.
* The second study used soy protein isolate with 90 mg isoflavones/day (48 mg genistein; 38 mg daidzein; 6 mg glycetin) compared to soy protein isolate without isoflavones and milk protein as a control. There was no significant effect on bone density amongst groups.
* A third study compared tablets containing 80 mg soy hypocotyl isoflavones/day (10 mg genistein; 44 mg daidzein; 27 mg glycetin) or 120 mg/day (15 mg genistein; 66 mg daidzein; 40 mg glycetin) to placebo. There was significantly less bone loss on total-body BMD with 120 mg/day and there was no effect on spine and hip. All three groups lost bone over the 2-year study period.
* And, the fourth study also compared tablets containing 80 mg soy protein isolate isoflavones/day (40 mg genistein; 31 mg daidzein; 9 mg glycetin) or 120 mg/day (60 mg genistein; 46 mg daidzein; 14 mg glycetin) to placebo. There was no significant effect on BMD among groups in the intent-to-treat analysis. After adjusting for age, body fat, and bone resorption, there was a significant effect of the 120-mg dose on femoral neck density although all three groups lost bone over the 3-year study period.
It is possible that the failure of isoflavones to prevent bone loss in the human studies is due to the low number of equol producers in the study groups.
* In a recent 1-year Japanese study of 75 mg isoflavones/day (38 mg daidzin; 0.6 mg daidzein; 8.6 mg genistin; 0.2 mg genistein; 24 mg glycitin with glycitein) compared to placebo, the results were analyzed according to equol status. There was no effect on spine BMD comparing equol producers and nonproducers but there was a significant benefit on total hip and femoral neck BMD in the equol producers.
* A 1-year double-blind RCT with 10 mg natural S(-)-equol supplements/day for 93 equol-nonproducing postmenopausal Japanese women resulted in some inhibition of urine bone resorption markers and significantly less bone loss at the hip though there was no effect on the spine. However, the effects are clearly less than those reported with low-dose estrogen. The potential bone benefits reported with natural S(-)-equol in Japanese women need confirmation in a Caucasian population.
Differences in soy products may also have different effects on bone metabolism. In a small study, the effect of soy cotyledon, soy germ, red clover, and kudzu were compared on calcium-41 excretion as markers of bone resorption and 1 mg estradiol/day was used as the active control. Serum genistein levels were four times higher on soy cotyledon than soy germ. Soy cotyledon and soy germ had a mild antiresorptive effect resulting in a decrease in bone resorption markers of 10% compared to 25% on estradiol.
There are no prospective intervention studies of soy isoflavones on the incidence of bone fracture. In an observational study of Chinese women over a 4.5-year period, women in the highest quintile of soy intake had a lower fracture rate. And there is some epidemiologic evidence that soy may reduce risk of osteoporotic fracture in women.
Conclusions
Despite the theoretical considerations suggesting that isoflavones might have efficacy on bone in human studies, the long-term studies are mainly negative, especially in view of the well-known beneficial effects of low-dose estrogen on bone. However, the possibility remains that the isoflavone content was not sufficiently high to bind to ER-β and inhibit bone resorption. An ongoing study in North America-the Women's Isoflavone Soy Health (WISH) Trial-is evaluating 25 g soy protein/day in postmenopausal women. The results from this study should be useful in the context of previous studies. Another limiting factor in soy studies might be that only 25% to 35% of women living in Western countries are equol producers.
Because there is currently no compelling evidence for the beneficial effect of soy isoflavones on bone density in postmenopausal women, more human studies need to be conducted with equol producers, with soy products with a higher genistein content, and with higher doses of isoflavones (>120 mg/d).