Dr. Messina is Adjunct Associate Professor, Loma Linda University, Loma Linda, C A, and President, Nutrition Matters, Inc., Port Townsend, WA. In the past three years, Dr. Messina has consulted for Archer Daniels Midland Company and the United Soybean Board.


Soy foods have played an important role in the diets of many Asian cultures for centuries. As they have made their way into Western diets, researchers have begun to study the claimed health benefits of soy-based foods — often with surprising results.

Media coverage of this research has sparked increased consumer interest in soy and skyrocketing sales of soy foods, which have more than tripled in the United States during the past decade. Mainstream food companies have jumped on the soy bandwagon and have developed "Westernized" versions of Asian soy products. Symbolic of the entrance of soy foods into the mainstream is the fact that about half of all soy products in the United States are sold in large retail markets, rather than in specialty stores.

Many of these new soy products are marketed specifically to women, in part because of evidence suggesting that soy may be helpful for conditions and diseases associated with menopause. Many view soy as an alternative to hormone replacement therapy (HRT) because it is a natural dietary source of isoflavones, a class of chemical compounds which affect the body in some ways similar to estrogen.

Soy may gain even further acceptance after the recently published findings of two studies: the Heart and Estrogen/Progestin Replacement Study (HERS) I/II, and the Women's Health Initiative (WHI). The results of both of these studies seriously call into question the safety and efficacy of routine, long-term use of HRT.

In the HERS I/II studies, HRT treatment failed to reduce the rate of heart attack and actually raised rates of dangerous blood clotting and biliary tract (gallbladder) surgery.1,2 Similarly, in the WHI study, health risks associated with HRT treatment so exceeded benefits that the trial was suspended after 5.2 years, instead of the planned 8.5 years. Given the current popularity of soy foods, it is important to consider whether what we have learned from the HERS I/II and WHI studies has any implications, negatively or positively, for soy and menopausal health.

What's So Special About Soy?

The soybean is a complex mixture of biologically active chemicals. The two components thought to be primarily responsible for any health benefits are protein and isoflavones. It is the isoflavones that largely account for the interest in soy as an alternative to HRT.

Isoflavones are a subclass of a large and common group of food substances called flavonoids. Unlike most flavonoids, isoflavones have a very limited distribution in the plant kingdom. The soybean is the only edible, naturally occurring dietary source of isoflavones (although isoflavones are now available in supplement form and are sometimes added to food).

While similar in quality to meat and milk protein, soy protein has two advantages: it appears to have a beneficial effect on kidney function and it is less likely to cause high levels of calcium in the urine in comparison to milk protein.

The technical names of the primary isoflavones in soybeans are genistein (4'5,7-trihydroxyisoflavone) and daidzein (4',7-dihydroxyisoflavone), and their respective b-glycosides, genistin and daidzein. Typically, more genist(e)in exists in soybeans and soy foods than daidz(e)in. There are also small amounts of a third isoflavone in soybeans, glycitein (7,4'-dihydroxy-6-methoxyisoflavone), and its glycoside, glycitin.

In the Spotlight: Isoflavones
More than 600 scientific papers are now published on isoflavones annually, compared with just 12 for the entire year of 1985. The US government is very interested in the health effects of isoflavones. For example, the National Cancer Institute (NCI) has been investigating the anticancer effects of soy, especially isoflavones, since 1991. In 1999, the US Department of Agriculture (USDA), in conjunction with Iowa State University, created an online database on the isoflavone content of foods, and in that same year, the National Institute of Aging (NIA) and the National Institute of Health (NIH) convened a three-day workshop on isoflavones.

As the chemical structure of isoflavones is similar to that of estrogen, it is not surprising that isoflavones can mimic estrogen's effect on the human body in many ways — but not all. This may be a good thing because estrogen has positive effects on some parts and systems of the body and negative effects on others.

The ideal estrogen substitute would have estrogen-like effects on the coronary vessels, skeletal system and brain but anti-estrogenic effects on the breast and uterus lining.

It appears that the isoflavones found in soy foods may do just that. Estrogen, for instance, increases endometrial cell proliferation and serum triglyceride levels — bad things from the point of view of disease risk — whereas isoflavone-rich soy protein — bad things from the point of view of disease risk — whereas isoflavone-rich soy protein and isolated isoflavones have no effect on endometrial cell proliferation in the uterus and little or no effect on serum triglyceride levels.

HERS I/II and WHI: HRT versus Soy

Breast Cancer
The WHI and HERS II studies indicated that breast cancer risk would be two to three times higher in lifetime HRT users. Several epidemiological studies have also found increases of this magnitude.

It seems, however, that it is the combination of estrogen and progestin, not estrogen alone, that increases breast cancer risk. Supporting this are several epidemiological studies which found that estrogen alone only slightly increases breast cancer risk, whereas the combination of the two hormones markedly increases risk. It is very important that, unlike the combination estrogen/progestin arm of the study referred to above, the estrogen-only arm of WHI was not discontinued and that, thus far, no increase in breast cancer risk has been reported.

Likely Effects of Soy
The first research on soy and breast cancer was undertaken not only because breast cancer mortality rates are generally low in Asia but also because weak estrogen-like compounds such as isoflavones, have been known for decades to exert anti-estrogenic effects under some experimental conditions In fact, the first study demonstrating the anti-estrogenic effects of genistein in rodents was published in 1966. Isoflavones can exert anti-estrogenic effects in several ways, although we have little direct evidence of these effects in humans. Overall, the evidence that soy intake by adults reduces breast cancer risk is not certain but an exciting hypothesis, based on studies of Japanese immigrants, as well as on animal and epidemiological studies involving soy is that soy consumption early in life markedly reduces adult breast cancer risk.

Despite the mounting evidence for the anticancer effects of soy, some researchers are concerned that the estrogen-like properties of isoflavones might increase breast cancer risk in certain high-risk women. However, although controversial, increasingly the available evidence suggests that neither soy nor isoflavones are a concern because the real culprit now appears to be the combination of estrogen and progestin, rather than estrogen alone. Clinical research also suggests soy is safe. In one study, isoflavone supplements had no effect on breast tissue density — increased breast density can be a precursor of breast cancer — in premenopausal women, whereas in another, isoflavones decreased breast tissue density in women 56-65 years of age. These effects are not only opposite to those of HRT, but suggest that soy either has no effect on or decreases breast cancer risk. The bottom line is that soy, unlike HRT, does not appear to increase breast cancer risk.

Venous Thromboembolic Disease
The WHI study found that HRT, and estrogen in particular, raises the risk of increased blood clotting, something that can cause diseases such as deep vein thrombosis and pulmonary embolism.
Likely Effects of Soy
Isoflavone-rich soy protein and isoflavones either reduce or have no effect on blood clotting. Animal research even suggests that soy and isoflavones may actually inhibit unhealthy blood clotting. Overall, there is no evidence to suggest that like estrogen, isoflavones, or soy, increase the risk of deep vein thrombosis or pulmonary embolism.

Stroke
According to the WHI study, HRT also increases the risk of stroke. Exactly how it does this is unknown, although one observational study reported that estrogen plus progestin is associated with a slightly higher risk of stroke than estrogen alone.

Likely Effects of Soy
There has been little investigation of the relationship between soy and stroke risk. However, one large epidemiological study of more than 18,000 Chinese men found that fatal stroke risk was unrelated to soy intake. Animal research suggests that genistein may be of value in the prevention of cerebral stroke via hormonal and nonhormonal mechanisms, but this is highly speculative

Studies of monkeys have reported beneficial effects from soy on stroke risk.

The incidence of stroke in Asia, where soy foods are commonly consumed, is relatively high compared with the West, but stroke rates in Asia have decreased dramatically over the past 30-50 years as hypertension has become less prevalent. The incidence of hemorrhagic stroke (the most common type of stroke in Asia) has declined markedly. A slight increase in serum cholesterol levels over the past several decades may also be a factor in the decline in intracerebral hemorrhage, as might be the increased intake of animal protein. Soy intake has remained fairly constant in Japan since about 1960, suggesting that soy does not increase stroke risk; furthermore, some evidence suggests that soy protein may decrease blood pressure.

Biliary Tract Surgery
In the HERS II study, women taking HRT were more likely to undergo biliary tract surgery.62 This effect probably results from estrogen because high-dose estrogen therapy has been shown to cause gallbladder disease. Estrogen may increase gallbladder disease risk by altering the concentration of cholesterol in the bile.

Likely Effects of Soy

Animal studies suggest that soy protein, in contrast to estrogen, may actually lower gallbladder risk by decreasing liver and biliary cholesterol concentrations.

Coronary Heart Disease
The WHI study found an increased risk of coronary heart disease (CHD), although there was no increased risk found in HERS II. This is somewhat puzzling, as previous studies had led researchers to expect a risk-lowering effect. Serum levels of LDL cholesterol decreased by 11% and 13%, and levels of HDL cholesterol increased by 10% and 7%, respectively, in the HERS and WHI. However, serum triglyceride levels, which may be an independent risk factor for CHD, increased by 8% and 7% in the HERS and WHI, respectively.

Likely Effects of Soy
In 1999, the US Food and Drug Administration approved a health claim for the cholesterol-lowering effects of soy protein and set 25 g/day as the target intake goal for cholesterol reduction. In 2000, the American Heart Association endorsed the use of soy foods for people with elevated cholesterol. The cholesterol-lowering effects of soy protein are relatively modest, particularly in people with only mildly elevated cholesterol, although there may be some individuals who are particularly sensitive to the cholesterol-lowering effects of soy protein.

Aside from cholesterol reduction, however, isoflavones may have other, more direct coronary benefits. They have been shown to enhance systemic arterial compliance, an indicator of arterial flexibility. Some evidence suggests that decreased arterial flexibility is associated with an increased risk of CHD. Isoflavone-rich soy protein may also decrease blood pressure and inhibit LDL cholesterol oxidation but the evidence for these effects is thin. Finally, soy and isoflavones may enhance arterial reactivity, which is thought to be a global indicator of coronary heart disease risk.

Only limited epidemiologic investigation of the relationship between CHD risk and soy intake has been conducted, but two Japanese epidemiologic studies supthe idea that soy can help keep the heart healthy.80,81 More importantly, a recent study from China found that soy intake was associated with an eighty percent reduction in the risk of having a non-fatal heart attack. Although no intervention studies to date have actually examined the impact of soy consumption on heart attack, the available evidence suggests that soy and isoflavones warrant inclusion in a regimen aimed at reducing CHD risk.

Colorectal Cancer
The WHI study found that HRT users were significantly less likely to develop colorectal cancer. Several epidemiologic studies have also found that HRT users are less likely to develop colon cancer and that estrogen may decrease colorectal cancer risk.

Likely Effects of Soy
Soy consumption is thought to fight certain forms of cancer and the evidence suggests that isoflavones are the primary anticarcinogens in soy. However, most of the research in this area has been on breast and prostate cancer. The evidence that soy reduces colon cancer risk is somewhat mixed. The most impressive study found that in subjects with a history of colon polyps or colon cancer who were fed 39 g soy protein/day for one year, there was a statistically significant decrease in colon cell proliferation and in the proliferation zone. These changes point to a major reduction in colon cancer risk.

If soy does reduce risk it is not clear whether isoflavones are responsible. At this point, the evidence is too limited to draw conclusions about the impact of soy or isoflavones on colon cancer risk.

Fractures
Not surprisingly considering the well-known ability of estrogen to prevent osteoporosis, WHI study found that HRT reduced a woman's risk of experiencing hip and vertebral fractures. These findings are generally consistent with other research.

Likely Effects of Soy
Because of the estrogen-like properties of isoflavones, there is much interest in their possible benefits on the skeletal system. Isolated isoflavones, as well as isoflavone-rich soy protein, inhibit bone loss in animal studies. Studies involving postmenopausal women are generally favorable, although a bit inconsistent and, as of today, positive effects have been noted primarily for the spine. However, most of these studies have been brief and small-scale. Furthermore, although there was considerable optimism about the skeletal benefits of the synthetic isoflavone ipriflavone, which is chemically similar to soybean isoflavones, ipriflavone recently failed to favorably affect bone density or fracture risk in a 3-year study involving approximately 500 women.

In addition to the short-term clinical studies, there is a fairly impressive amount of epidemiologic data indicating that, among Asians, those women who consume above average amounts of soy or isoflavones have higher bone density, though not all studies are in agreement. Finally, in addition to the possible benefits of isoflavones, soy protein, when substituted for animal protein, may improve bone health, retaining the body's calcium needed for bones by cutting down the amount of calcium excreted in the urine. Evidence suggests that every gram of soy protein substituted for dairy protein decreases urinary calcium excretion by 0.5-1.0 mg/day. With all other things being equal, this level of substitution over the course of many years is likely have significant health benefits.

Conclusions
Much less is known about the health effects of eating soy than about those of HRT. So far, no studies have actually examined the impact of soy consumption on disease outcomes; only markers of disease risk have been considered. As demonstrated by the results of HERS I/II and WHI, long-term intervention studies that focus on actual disease outcomes are necessary before we can draw definitive conclusions about the health effects of any biologically active agent. Therefore, making predictions based on the results of these trials to soy should be done very cautiously.

The HRT preparations used in HERS I/II and WHI consist not of a single estrogen but of a complex mixture of estrogens derived from the urine of pregnant mares, in combination with a synthetic progestin. Obviously, this chemical mix is very different from soy and isoflavones. Nevertheless, on the basis of the available data, there is little reason to think that soy consumption will increase any of the disease risks that were associated with HRT in HERS I/II and WHI. There is, however, at least preliminary data to suggest that soy may decrease colorectal cancer and fracture risk, but this remains speculative.

Therefore, if the results of HERS I/II or WHI have implications for soy, they are that soy may have many of the advantages and few of the disadvantages of HRT, but that large, long-term intervention studies are needed before definitive conclusions can be drawn. Whether soy alone can serve as a safe and effective alternative to HRT remains to be determined. There is no doubt, however, that menopausal women should be encouraged to consume isoflavone-rich soy products, although the effect of soy or isoflavone supplements on the alleviation of hot flashes is less pronounced than that of estrogen. Also, a recent analysis concluded that soy is effective primarily in women with frequent hot flashes.

Soy Intake Recommendations
In approving a health claim for soy protein, the US Food and Drug Administration set 25 g soy protein/day as the target intake goal for cholesterol reduction. However, because the FDA was focusing on those with high cholesterol levels, this level of intake is too high to be used as a reference point for all healthy adults. Twenty-five grams of soy protein/day is about half the recommended daily allowance for protein for adult women and is approximately two times the average soy protein intake of Japanese adults. A more appropriate intake recommendation for the average person is 15 g (with a range of 10-25 g) soy protein and 50 mg (aglycone weight) isoflavones (range of 30-100 mg/day).

Traditional soy foods have an isoflavone (mg) to protein (g) ratio of approximately 3.5:1; therefore, consuming 15 g soy protein will result in consuming approximately 50 mg isoflavones. These amounts of soy protein and isoflavones are proughly equivalent to two servings of traditional soy foods and are likely to be effective for those diseases for which soy is proven to be beneficial. In fact, this amount of soy protein may even help lower cholesterol levels, as shown by recent studies Furthermore, even those with traditional Western eating habits can easily incorporate 15 g soy protein into their daily diets. As this recommendation is consistent with the universal advice of experts to eat a varied diet, there is likely to be little down side for most people.

One final perspective by which to view this recommendation is to consider that substituting 15 g soy protein for animal protein would cause the current US dietary animal-to-plant protein ratio to fall from 2:1 to 1:1, the ratio of the US diet during the early 1900s. And, yet, soy protein would still represent less than 20% of the typical protein intake of US adults. Thus, soy could serve to bring more balance to the US diet and, at the very least, would function as a source of high-quality protein that lacks the saturated fat and cholesterol typical of the forms of protein most commonly eaten by US residents.