Clare M. Hasler, Ph.D., M.B.A

Safety of Soy Protein Isolates

“Cinderella’s Dark Side” is a feature story on Dr. Joseph Mercola’s anti-soy website( which lists a myriad of reasons whywe should not consume soy. On this site, soy protein isolate (SPI) is deemed to be “notso friendly” because it “is not something you can make in your own kitchen” (how thisrelates to the so-called “danger” of soy is completely unclear) and the fact that it containsnumerous anti-nutrients such as trypsin inhibitors and goitrogens. This story also statesthat the high temperature processing of SPI to remove trypsin inhibitors has theunfortunate side effect of “so denaturing the other proteins in soy that they are renderedlargely ineffective” suggesting that soy is a poor quality protein source. This synopsiswill briefly review the safety and quality of soy with an emphasis on soy protein isolate.

Protein quality. Proteins are an essential component of the diet needed for the survivalof animals and humans. Soybeans are a major source of high quality vegetable protein inmany countries around the world. Protein constitutes approximately 40% of the total drymatter of soybeans, making this essential nutrient the component present in soybeans inthe greatest amount. Soy protein provides calories, nitrogen, as well as essential aminoacids-the building blocks of protein. Although there are hundreds of amino acids innature, only approximately 20 appear in proteins and only nine of these are “essential”which means we need to get them from dietary sources because our bodies do notsynthesize them. The essential amino acids include: threonine, cyst(e)ine + methionine,valine, isoleucine, leucine, tyrosine + phenylalanine, histidine, lysine, and tryptophan.

The nutritional quality of proteins is largely dependent on the pattern of amino acids in aparticular protein source. More specifically, protein quality refers to how closely afood’s essential amino acid (EAA) pattern matches the needs of the body. Plant proteinstend to be limiting in one or more of the EAAs. Like proteins of most other leguminousplants, soy protein is low in sulfur-containing amino acids, with methionine being themost limiting, followed by cyst(e)ine and threonine (Eggum and Beames, 1983). Avariety of chemical methods, in vivo assays (using either animals or humans), and in vitroacids methods (employing various proteinases) have been developed to evaluate proteinquality. Since 1919, the preferred method of evaluating protein quality in both the U.S.and Canada has been the Protein Efficiency Ratio (PER). However, this method is basedon the growth of young rats and since rats have a higher relative requirement for sulfur-containing amino acids (to support fur growth), the PER method tends to undervalue theprotein quality of soybeans for humans (Liu, 1999). Another factor important in thedetermination of protein quality is digestibility, which is an index of bioavailability, andis not included in certain methods of assessing protein quality (e.g., amino acid score). To overcome the limitations of older methods of protein quality evaluation, a newmethod of evaluating protein quality was adopted by the World Health Organization in1990 (FAO/WHO, 1990) called Protein Digestibility Corrected Amino Acid Score(PDCAAS). digestibility, and (3) ability to supply EAA in amounts required by humans and iscalculated as follows:


Figure 1.

The highest possible score using PDCAAS is 1.0. All proteins with a PDCAAS of 1.0are considered equally high in quality and provide all of the EAAs. According to theFAO/WHO (1990), SPI has a PDCAAS of 1.00, which is equivalent to casein and eggwhite protein. Thus, based on the PDCAAS method, the protein value of well-processedSPI is essentially equivalent to that of animal proteins and thus can serve as the major, oreven sole, source of protein intake (Young, 1991).

Soy anti-nutrients. A basic principle of toxicology is that all chemicals are “toxicchemicals.” It is the dose that makes the poison. This tenet holds true for all foods andfood components, including soy and soy phytochemicals. More specifically, all foodscontain components that, when consumed in excessive amounts, produce unwanted sideeffects—including the fruits and vegetables that we consume everyday and that arecurrently promoted for reducing cancer risk! An excellent example was eloquentlypresented in a classic paper published in the Proceedings of the National Academy ofSciences, one of the most prestigious biomedical journals, by Professor Bruce N. Ames ofthe University of California at Berkley. The review compared synthetic chemicals suchas dioxin to natural chemicals, such as those found in broccoli and cabbage anddetermined that in high-dose tests, a high proportion of both natural and syntheticchemicals are carcinogens, mutagens, teratogens, and clastogens (i.e., agents that causeDNA breakage). In fact, about one-half of the natural “pesticides” that exist in manyfruits and vegetables--including cabbage, grapefruit and broccoli--are cancer-causingagents at roughly the same proportion as synthetic pesticides (Ames et al., 1990). Natural toxins have the same mechanisms of toxicity as synthetic toxins. Soybeanscontain a broad spectrum of 136 physiologically active components (Fang et al., 2004).Some of these components have been termed “anti-nutrients,” which are defined ascompounds that hinder the utilization of one or more nutrients and/or adversely affectnutritional status and health. The protease inhibitors have been considered to be a soy-antinutrient.

Protease inhibitors. Two primary types of protease inhibitors are present in soybeans: (1) the Kunitz inhibitor and (2) the Bowman Birk Inhibitor (BBI). The former inhibitsthe activity of the enzyme trypsin and the latter inhibits both trypsin and chymotrypsin.Approximately 6% of soybean protein is comprised of protease inhibitors (Rackis andAnderson, 1964) and the nutritional significance of this bioactive component has been thesubject of considerable debate for many years (Leiner, 1995). Although proteaseinhibitors have been shown to suppress growth in young animals (Birk, 1993) as well aslead to hypertrophy and hyperplasia of the pancreas and pancreatic cancer at high levels (McGuiness et al., 1984). There is absolutely no evidence that these effects occur inhumans. Further, protease inhibitors are heat labile and about 80% of the trypsin activityis destroyed during the commercial processing of soybeans into soyfoods (Rackis andGumbmann, 1982). Moreover, considerable research has demonstrated that BBI mayhave anti-carcinogenic activity (Kennedy, 1994). Protease inhibitors have been shown tohave cancer chemopreventive activities both in vitro and in vivo (Kennedy, 1993). Inanimal carcinogenesis studies, a concentrate of BBI, BBIC, has been shown to suppresscarcinogenesis in a wide variety of in vivo models of various types of cancer, includingcolon, lung, liver, oral, and esophageal (Kennedy, 1995). Human cancer prevention trialswith BBI are ongoing.

Data from human studies suggest that dietary soy or isoflavones are unlikely to have anadverse effect on thyroid function in normal individuals with adequate iodine intake.However, it is conceivable that the thyroid function of hypothyroid individualsconsuming high levels of isoflavone supplements may be adversely affected. Further,isoflavones could potentially interact with thyroxine medication in individuals diagnosedwith congenital hypothyroidism and could lower the amount of thyroxine available in thefree (active) form.

The thyroid gland is responsible for the production of hormones involved in regulatingmetabolism, body weight and oxygen requirements as well as normal growth anddevelopment during childhood. There are two primary thyroid hormones: (1) T3 (tri-iodothyronine) and T4 (thyroxine), both of which are synthesized in the thyroid glandfrom iodine and the amino acid tyrosine. The amount of T3 and T4 produced by thethyroid gland is controlled by thyroid stimulating hormone (TSH), which is secreted fromthe pituitary gland and is regulated by the central nervous system.

Goitrogens. It was first reported in the 1930’s that soybeans had “goitrogenic” activity inrats (McCarrison, 1933), that is, caused the development of a goiter (an enlargement ofthe thyroid gland). Since then, other studies have shown that dietary soy or isoflavonescan affect the thyroid function of rodents (Balmir et al., 1996; Mitsuma et al., 1998; Ikedaet al., 2000; Son et al., 2001). This is because isoflavones have a similar structure to T3and T4, and studies conducted in vitro demonstrate that the isoflavones inhibitthyroperoxidase (TPO), an enzyme involved in the synthesis of T3 and T4. However,although reductions in TPO have been seen in rats fed isoflavones, the remaining activityof this enzyme is sufficient to maintain normal thyroid homeostasis. Further, Chang andDoerge (2000) found no differences in T3, T4, TSH concentrations or thyroid glandweight or histopathology in rats continuously fed a soy diet containing 60 mggenistein/kg diet compared with control animals.

Several studies have examined the effect of consumption of soybeans or isoflavones onthyroid function in adults (Duncan et al., 1999a; Duncan et al., 1999b; Persky et al.,2002; Jayagopal et al., 2002). Overall, data from these studies suggest that dietary soy orisoflavones are unlikely to affect thyroid function in normal individuals with adequateiodine intake. However, isoflavones could potentially interact with thyroxinemedication given to patients diagnosed with congenital hypothyroidism. If a fixed dose of thyroxine is used in treatment, ingestion of large amounts of soy isoflavones couldlower the amount of thyroxine available in the free (active) form.

Very few studies have investigated the possible associations between soy isoflavones andthyroid cancer. A recent retrospective case-control study involving 1166 subjects in SanFrancisco (Horn-Ross et al, 2002) which examined the relationship between isoflavoneconsumption and risk of thyroid cancer actually found that increased consumption ofunfermented soy-based foods was associated with a decreased risk of developing thyroidcancer. Tthere is no evidence that soy has an adverse effect on the thyroid gland innormal individuals with adequate iodine intake. Further, populations that consumerelatively high amounts of soy (e.g., Japan) do not have a significantly higher incidenceof hypothyroidism.

In summary, research continues to demonstrate the healthfulness and safety of soyfoodsat doses currently recommended by the Food and Drug Administration, the AmericanHeart Association and other public health organizations. Healthy adults should strive toconsume approximately 10 to 25 g/day of soy protein for optimal health. This is not onlya practical level of intake, but is also based on the level of soy intake observed in Asiancountries (where there is a reduced risk of age-related chronic diseases) as well asevidence from safety and efficacy studies.

Soy and Pregnancy

A high estrogenic environment in utero may increase subsequent breast cancer riskaccording to a 1999 study in Oncology Reports highlighted on Dr. Joseph Mercola’swebsite ( Theresults from this study indicate that in utero exposure to genistein, dose-dependentlyincreased the incidence of breast tumors when compared with controls (Hilakivi-Clarkeet al, 1999a). The take home message is, according to the website: “avoid soy,especially if you are a pregnant woman.”

This recommendation is simply ludicrous and taken completely out of context as is mostinformation on this website.

Although the study by Hilakivi-Clarke et al. suggests that in utero exposure to genisteinmay increase the incidence of mammary tumors in the offspring, several additionalstudies are in direct conflict with these findings and support exactly the oppositehypothesis: that in utero exposure to soy isoflavones and genistein in particular mayactually reduce breast cancer incidence later in life (Lamartiniere et al., 2000). Inaddition, another study published the same year by Hilakivi-Clarke et al. (1999b)reported that prepubertal exposure of rats to dietary genistein (1 mg/kg body weight/day)decreased the number of tumors per animals as well as tumor growth. Fritz et al. (1998)also showed that perinatal exposure to genistein (from conception to post natal day 21)through the maternal diet (25 or 250 mg genistein/kg diet) resulted in a dose-relateddecrease in chemically induced mammary tumors in rats. Most importantly, there is noevidence from human epidemiological studies that soy consumption during pregnancyincreases the risk of breast cancer.

Exposure to phytoestrogens during development or early life may play an important rolein programming hormonal homeostasis and thus influence an individual’s later life risk ofdeveloping cancer. Although the animal data on breast cancer and exposure toisoflavones is somewhat conflicting (as is the case in most areas of research), a number ofstudies have shown that genistein has a protective effect in animal models of chemicallyinduced cancer.

In May of 2003, a comprehensive report of the Committee on Toxicology of Chemicalsin Food, Consumer Products and the Environment was drafted to advise on the healthimplications of dietary phytoestrogens through a comprehensive review of publishedscientific research and the United Kingdom’s Food Standards Agency’s PhytoestrogenResearch Programme. This 441 page report concluded: “animal studies suggest thatexposure to phytoestrogens in early life inhibits development of breast cancer later inlife. ” Although this working group recommends that future research examine whateffects maternal exposure to phytoestrogens may have on the fetus and on the subsequenthealth status of the child (because there are no human studies published on this topic),nowhere in this report does it state that pregnant women should avoid soy duringpregnancy.

There is ample reason to think that research currently underway will show that soyconsumption is not contraindicated for women in any situation. In the meantime,pregnant women should feel confident (and in fact should consider) consuming amoderate amount of soy (10-25 grams) and soy isoflavones (30-100 mg) every day(Messina and Messina, 2003) for optimal health.

Soy Infant Formula: A Safe Viable Feeding Option for Infants

“All soy formula is worse than worthless for human infants and is nearly guaranteed tocause problems down the road” according to Dr. Joseph Mercola(; accessed June 26, 2004). Notsurprisingly, this is the same “expert” who states: “it is important to note that whenbreast feeding it is wise to avoid drinking milk...” No one would argue that “breast isbest” when it comes to infant feeding. The American Academy of Pediatrics iscommitted to the use of maternal breast milk as the ideal source of nutrition for infantfeeding. However, by 2 months of age, most infants in North American are formula-fed;soy-based infant formula now constitute 25% of infant formula sales. Consequently, thenutritional adequacy and safety of soy protein is of paramount importance to thisvulnerable segment of the population. Soy protein-based nutrition has been used duringinfancy for centuries in the orient and experts agree it is a perfectly safe feeding optionfor infants. This synopsis will correctly present the facts on soy infant formula as regardsthe presence of phytoestrogens, impaired thyroid function, and manganese and aluminumconcentrations.


Manganese is a trace element essential for life. Manganese toxicity isextremely rare, but has been reported in mine workers exposed to high concentrations ofmanganese dust, resulting in a disorder known as “manganese madness” (PDR forNutritional Supplements, 2001). There are also a few reports of manganese intoxicationoccurring in those on long-term total parenteral nutrition (TPN). Dietary or supplementalforms of manganese are quite safe. According to the Food and Nutrition Board of theU.S. National Academy of Sciences, the estimated safe and adequate daily dietary intake(ESADDI) level of manganese for infants aged 0 to 6 months is 300 to 600 mg; up to 10mg is considered safe (PDR for Nutritional Supplements, 2001). There have been noreports of manganese toxicity in healthy infants fed soy-based formulas. Manganeselevels in soy-based infant formula are higher than that of human breast milk. Humanbreast milk contains 1 microgram of manganese per 100 ml while soy-based formulascontain 25 micrograms of manganese per 100 ml. Thus, manganese levels in soy formulaare more that of breast milk. Although one group of researchers did report neurotoxicityin rats given 500 micrograms of supplemental manganese (not soy formula) per day(Tran et al., 2002), this level of manganese would be impossible to attain by an infantconsuming soy-based formula. In addition, rats absorb manganese differently thanhuman infants and are more prone to manganese toxicity. There have been nopublished studies linking the manganese content in soy formula fed to healthyinfants to any adverse effect.


Aluminum, the third most common element after oxygen and silicon, is notconsidered an essential nutrient for humans. However, it is widespread in food and watersupplies because of its presence in soil, water and air. Due to the fact that the soybeanplant accumulates aluminum from the soil, soy-based infant formulas are known tocontain high levels of aluminum. The aluminum content of human milk is 4 to 65 ng/mL,while that of soy protein-based formula is 600 to 1300 ng/mL (Fomon and Ziegler, 1979).Despite the higher aluminum concentrations in soy formula, serum aluminum levels inbreast fed infants do not differ significantly from levels in infants fed soy formula (Litovet al., 1989). Although there was a case report published in the Lancet in 1985documenting a high concentration of aluminum in the brain of two infants withcongenital kidney disease associated with the consumption of soy formula (Freundlich etal., 1985), these same researchers later acknowledged that unrecognizable sources ofaluminum (e.g., intravenous fluids) may have contributed to the excessive concentrationsof aluminum in the brain (Freundlich et al., 1990). No other case reports have foundproblems with aluminum in soy formula. Aluminum from infant formula is not of concern for infants with normal kidney function, since the kidney absorbs very littlealuminum and that which is absorbed is excreted by the kidney and eliminated throughthe urine. Further, the aluminum intake of infants using soy formula is only about 25%that of the upper tolerable level established by the Food and Agriculture Organization. The aluminum content of soy formula is not viewed as a contraindication to its use.

Thyroid Function

There is no convincing evidence that soy protein has an adverseeffect on thyroid function in healthy human infants consuming adequate iodine (Messina,2001). There is evidence that animals exposed to large amounts of soy protein willdevelop goiter, particularly when fed an iodine deficient diet (Filisetti and Lajolo, 1981).This is due to the fact that the principal isoflavones in soy, genistein and daidzein, havebeen shown to inhibit thyroid peroxidase (Divi et al., 1997) and 5'-deiodinase (Cody etal., 1989), key enzymes involved in thyroid hormone biosynthesis. The inhibition of theseenzymes results in decreased levels of circulating thyroid hormones (e.g., T4 and T3),which leads to increased secretion of thyroid stimulating hormone (TSH) by the anteriorpituitary. The increased levels of TSH provide a growth stimulus to the thyroid, resultingin goiter. It must be emphasized, however, that this occurs only with very large amountsof soy isoflavones in the diet and/or when the diet is low in iodine. Researchers at theNational Center for Toxicological Research in Jefferson, Arkansas have found that, eventhough substantial amounts of thyroid peroxidase activity are lost when soy isoflavonesare consumed by normal rats, the remaining enzymatic activity is sufficient to maintainthryoid homeostasis in the absence of additional perturbations (Chang and Doerge, 2000).Further, dietary soy isoflavones are not the only dietary flavonoids that can inhibitthyroid peroxidase. A variety of other flavonoids have also been shown to be even morepotent in inhibiting the activity of this enzyme, including kaempferol, naringenin, andquercetin (Divi and Doerge, 1996). Such flavonoids are widely distributed in plant-derived foods and would be consumed daily at relatively high levels (possibly up to 1gram or more per day) by vegetarians or semi-vegetarians, yet these individuals do nothave a significant increased incidence of goiter. In the late 1950s, 10-15 cases of goiterwere identified in infants fed non-iodized soy flour-based infant formula. However, thistype of formula has not been used since the 1960s. Today, soy formula is based on soyprotein isolate and is fortified with iodine. No cases of goiter in infants, due to theconsumption of soy protein isolate-based iodized formula as is used today, have beenreported in the scientific literature.


Soy infant formula has been used safely by millions of infants over the course of the lastseveral decades. According to a 1998 policy statement from the American Academy ofPediatrics: “In term infants whose nutritional needs are not being met from maternalbreast milk of cow milk-based formulas, isolated soy protein-based formulas are safeand effective alternatives to provide appropriate nutrition for normal growth anddevelopment” (American Academy of Pediatrics, 1998).

Soy and Breast Cancer: What’s the Real Story?

The low breast cancer mortality rates in soyfood consuming populations has promptedresearchers to examine the role of soy and soy components in reducing the risk of varioustypes of hormone-dependent cancers (Messina et al., 1994; Fournier et al., 1998),particularly breast cancer (Barnes et al., 1997). The relationship between soy intake andbreast cancer risk is one of the most controversial areas of soy research today. This isprimarily because soy contains isoflavones, which have estrogen-like activity and greaterlifetime exposure to estrogen has been associated with increased breast cancer risk.Further, a small number of animal studies resulting from one research group have shownthat when immune deficient mice who have had their ovaries removed (to stop estrogenproduction) were implanted with estrogen receptor positive breast cancer cells and thengiven the isoflavone genistein, tumor growth was stimulated in comparison to mice notgiven genistein (Hsieh et al., 1998; Allred et al., 2001). However, this model of breastcancer has been very highly criticized on methodological grounds and cannot beextrapolated to postmenopausal women who have some (albeit low) level of circulatingestrogen. Further in a similarly designed experiment in which mice were notovariectomized (and thus serum estrogen levels were high), genistein actually inhibited,rather than stimulated tumor growth (Shao et al., 1998). Nevertheless some so-calledexperts state that that soy consumption may actually increase the risk of breast cancer( accessed August 2, 2004).However, there is absolutely no convincing epidemiological or clinical evidence that soyconsumption increases a woman’s risk for breast cancer. In fact, some experts state thateven breast cancer patients can safely consume moderate amounts of soy (Messina andLoprinzi, 2001). This review will briefly review the role of soy in breast cancer risk andpresent a balanced view of the evidence.

Interest in the potential role of soy in breast cancer prevention stemmed initially from theobservation that women from Southeast Asia have very low rates of breast cancermortality compared to Western women. For example, the breast cancer death rates (per100,000) in Japan are 6.7 compared to 27.4 in the U.S. (Meng et al., 1997). In addition,when women from Southeast Asia immigrate to Western countries, their breast cancerrates increase within one to two generations, suggesting that differences in breast cancerrates between Southeast Asian and Western countries are not due to genetics but ratherenvironmental factors (Shimizu et al., 1991), including dietary. Lending further credenceto this hypothesis is the fact that soy intake in Southeast Asian populations issignificantly higher than that of Western populations (Nagata, 2000). These observationsprompted the National Cancer Institute to hold a workshop on the potential role of soy inreducing the risk of cancer, during which five known anti-carcinogenic compounds insoybeans were identified: saponins, phytates, protease inhibitors, phytosterols, andisoflavones (Messina and Barnes, 1991). Of these, it has been unquestionably theisoflavones that have garnered the most research attention. Over 600 research papers arepublished annually on this soy phytochemical (Lu et al., 2001).

Isoflavones are a type of flavonoid, a class of compounds widely distributed throughnature. However, soybeans are the only significant dietary source of these bioactivecomponents as demonstrated in the United States Department of Agriculture’s onlinedatabase: ( From abiochemical and physiological standpoint, isoflavones are intriguing in that they have amolecular structure similar to that of the human estrogens and thus elicit estrogen-bindingactivity. However, isoflavones act as weak estrogens, binding to the estrogen receptorswith only 10-5 and 10-2 of the activity of 17β-estradiol. Additionally, more recentresearch has demonstrated that isoflavones bind to, and preferentially activate, the newlydiscovered estrogen receptor, estrogen receptor beta (ER-B), which is differentiallyexpressed in tissues. Isoflavones are increasingly being viewed as selective estrogenreceptor modulators (SERMS), meaning that they exert estrogen-like effects on thebones, coronary vessels and the brain and anti-estrogenic effects on the breast andendometrial tissue (Setchell, 1998). In fact, some experts object to referring toisoflavones as phytoestrogens. SERMs such as tamoxifen (a breast cancer therapeuticagent) and raloxifene (an osteoporosis therapeutic agent) have estrogen-like effects insome tissues but either null or antiestrogenic effects in other tissues. Similarly, soyisoflavones are thought to exert the same beneficial effects of estrogen without thedisadvantages, including increasing cancer risk.

With the exception of the studies utilizing athymic, immune-deficient mice discussedabove, the bulk of in vivo studies show that soy at least modestly inhibits mammarytumorigenesis in adult animals. Although the addition of soy to a standard laboratory dietdoes not significantly inhibit tumor incidence (the percentage of animals in the groupwith tumors, in most cases, soy consumption does inhibit tumor multiplicity (number oftumors per animal) by 25 to 50% (Haddak et al., 2000). Perhaps the most intriguinganimal data come from researchers at the University of Alabama (Lamartiniere, 2000).They have shown that in rats, neonatal exposure to genistein reduces later development ofcarcinogen-induced mammary cancer by approximately 50% (Lamartiniere et al., 1995a;Lamartiniere et al., 1995b; Murrill et al, 1996). These findings are thought to be due tothe fact that exposure to genistein during critical periods of mammary gland developmentcan render the mammary gland less susceptible to DNA damage by carcinogens later inlife. These animal studies are supported by recent epidemiological findings. In arecently conducted large scale case-control study in Shanghai (Shu et al., 2001), womenwho consumed approximately 11 grams of soy protein per day during their teenage years(13-15) were almost 50% less likely to develop breast cancer as adults than adult womenwho consumed ≤2 grams of soy protein/day during this period. Thus, it may beimportant for adolescent and teenage girls to consume soy daily at a level resemblingAsian soy consumption (15 grams of soy protein; 50 mg isoflavones) to reduce risk ofbreast cancer later in life (Messina and Messina, 2003).

Increases in mammographic density have been associated with a 4 to 6-fold increasedrisk of breast cancer (Atkinson et al., 1999) and soy intake has been associated withreduced mammographic density patterns. A randomized, placebo controlled studyinvestigating the effect of an isoflavone supplement (40 mg/day) has suggested asignificant reduction in density in women aged 56-65 compared to age matched controls

(Atkinson and Bingham, 2002). Similar results were noted in a cross-sectional study inSingapore-Chinese women who were asked to self-report dietary intake of soy and soyisoflavones (Jakes et al., 2002). Women with the highest reported dietary intake of soyand soy isoflavones were associated with low-risk mammographic parenchymal patterns.

Although two human studies have prompted concerns about women with estrogenreceptor positive breast cancer consuming soy, both of these studies weremethodologically flawed. The first study found that daily consumption of 38 g soyprotein over 5 months in premenopausal women was associated with an increase in breastnipple aspirate fluid secretion and breast cell proliferation (Petrakis et al., 1996), which istypically viewed as a marker for increased breast cancer risk (Preston-Martin et al.,1993), but not always (Mommers et al., 1999). However, this study lacked a controlgroup and fluid secretion also continued to increase in women even after soy feeding wasdiscontinued. Further, Bouker and Hilakivi-Clarke (2000) noted that women wereeligible for the study only if they were secretors of nipple aspirate fluid. The secondstudy that has raised concern in breast cancer survivors examined the effects of feeding60 g of textured vegetable protein (containing 45 mg isoflavones) for 2 wk on breast cellproliferation in premenopausal women with benign or malignant breast disease Apreliminary analysis of this study based on biopsies from only half of the subjectsindicated soy consumption markedly increased breast cell proliferation (McMichael-Phillips et al., 1998). However, in the final analysis, which included all 84 subjects, nosuch effect on breast cell proliferation was noted (Hargreaves et al., 1999).

In conclusion, on the whole, the evidence suggests that consuming moderate amounts ofsoy is much more likely to be of overall benefit to health rather than harmful, both interms of breast cancer risk and other chronic diseases.

Soy and Menopause

Menopause is defined as the spontaneous, permanent ending of menstruation that is notcaused by any medical intervention. In the Western world, most women experiencenatural menopause between the ages of 40 and 58 with the average age being 51 (NorthAmerican Menopause Society, 2003). Perimenopause includes the 3-6 year intervalbefore the last period and is characterized by wildly fluctuating hormones resulting in aplethora of symptoms, including hot flashes, night sweats, vaginal dryness, insomnia andmood swings (Love, 2003). The hot flash is the most common discomfort experiencedby perimenopausal women.

The potential role for soy and/or soy isoflavones as an alternative for hormonereplacement therapy (HRT) (Messina, 2003; Wuttke et al., 2003) and particularly in thealleviation of the vasomotor symptoms associated with menopause (Messina and Hughes,2003) has been a subject of much discussion over the last several years, particularly inlight of the results of two large clinical trials: (1) Heart and Estrogen/ProgestinReplacement Study and (2) Women’s Health Initiative (WHI)—both of which showedthat the long-term possible harm of HRT outweighed any potential benefit.

The WHI was a National Institutes of Health (NIH) multi-center trial that began in 1993.One arm of the trial consisted of a randomized, blinded, placebo-controlled hormonestudy involving 16,608 women, aged 50 to 79 (average age 63.2) who received eitherplacebo or continuous combined estrogen (0.625 mg/day conjugated equine estrogens)-progestogen (2.5 mg/day of medroxy-progesterone acetate) therapy (Prempro). Risk ofcoronary heart disease (CHD), stroke and breast cancer increased by 29%, 41% and26%, respectively, while the risk of colon cancer and osteoporosis of the hip and spinedecreased 37%, 34% and 34%, respectively (Rossouw et al., 2002). Because of thesefindings, the combined estrogen-progestogen arm of the trial was terminated after 5.6years rather than the planned 8.5 years. Although the estrogen-only treatment group wasallowed to continue, this arm was also terminated on February 2, 2004 when the NIHconcluded that, after an average of nearly 7 years of follow-up, estrogen alone does notappear to affect (either increase or decrease) heart disease, a key question of the study.At the same time, estrogen alone appears to increase the risk of stroke similar to whatwas found in the WHI study of estrogen + progestin when that trial was stopped in July2002. (, accessed August7, 2004). No increase in breast cancer was noted. As HRT data have been furtheranalyzed the cardiovascular and dementia risks have also been identified. In summary,according to WebMD Health ( accessed August 7, 2004):

  • HRT-related breast cancers first become apparent after 4 years of HRT use. Thenumber of HRT-related breast cancers increased with each additional year ofHRT use. Women taking HRT generally had larger, more advanced tumors thanwomen who developed breast cancer while taking placebo treatment (Rossouw etal., 2002).
  • HRT slightly increases heart attack and stroke risk in all healthy postmenopausalwomen regardless of risk factors (Manson et al., 2003).
  • HRT slightly increases the risk of blood clots in the lungs and legs in all healthypostmenopausal women regardless of risk factors (Wassertheir-Smoller, 2003).
  • HRT increases the risk for Alzheimer’s disease and other dementias in womenaged 65 and older. The increased risk first becomes apparent in women takingHRT for more than 4 years. The WHI researchers have concluded that HRT doesnot provide protection from dementia or cognitive impairment, as was previouslybelieved (Shumaker et al., 2003).
  • Among HRT users, the number of abnormal mammograms increases byapproximately 4% per year, first apparent after 1 year of HRT use. Daily estrogenplus progestin increased breast density compared to estrogen alone or placebo.

The Heart and Estrogen/progestin Replacement Study (HERS), was a randomized clinicaltrial of estrogen plus progestin, with or without statin drugs, vs. placebo, in 2763postmenopausal women with heart disease. HRT resulted in a statistically significantincrease in early risk for primary events in women who did not use statins (RelativeHazard =1.75, 95% CI 1.02 to 3.03, P=0.04) but not in statin users (RH=1.34, 95% CI 0.63 to 2.86, P=0.45).

Due to findings from the WHI and HERS studies, the percentage of women aged 50 to 74taking HRT has declined from 42% in 2001 to 38% in July 2003. The decline in the useof Prempro, the specific type of HRT used in the WHI has been even more dramatic—decreasing by 70%. Women are seeking alternatives to HRT and one of the leadingalternatives they are turning to is soy.

The role of soy in the amelioration of hot flashes and night sweats associated withmenopause was first noted more than 10 years ago by Lock (1991) in a survey of over2,600 Japanese and Canadian women. She noted that 30.9% of the Canadian women hadexperienced a hot flash in the preceding two weeks compared to only 9.7% of theJapanese women. In addition, only 3.6% of the Asian women experienced night sweatscompared to 19.6% of the Canadian women. These findings are consistent with the factthat approximately two-thirds of North American women experience perimenopausal hotflashes (North American Menopause Society, 2003) while American women of Chineseand Japanese ancestry are about one-third less likely to report experiencing hot flashes(Gold et al., 2000).

It was first suggested by Adlercreutz et al. (1992) that the estrogen-like properties ofisoflavones might explain the low incidence of hot flashes reported by women in Japan.Since that time, more than two-dozen clinical trials have been conducted to investigatethe efficacy of either soy or red clover isoflavones in alleviating hot flashes. A recentreview of 19 trials involving more than 1,700 women evaluated the efficacy of soyfoodsand isoflavones supplements for the alleviation of hot flashes (Messina and Hughes,2003). Overall, they found that there was a statistically significant relationship (p=0.01)between initial hot flash frequency and treatment efficacy with initial hot flash efficacyexplaining about 46% of the treatment effects. The frequency of hot flashes decreased byapproximately 5% (above placebo or control effects) for every additional hot flash efficacyexplaining about 46% of the treatment effects. The frequency of hot flashes decreased byapproximately 5% (above placebo or control effects) for every additional hot flash perday in women who experienced five or more hot flashes per day. More specifically, ofthe 11 studies that examined the effects of soyfoods (see references 48-58 in the review by Messina and Hughes), only one (Albertazzi et al., 1998) found a statisticallysignificant decrease in hot flash frequency in the treatment versus the control group.Although an additional study by Washburn et al. (1999) found that soy protein isolatereduced hot flash severity but not incidence, this only occurred when the isolate wasconsumed twice per day.

In contrast to the relative lack of effect of soy foods on hot flashes, four out of fivestudies utilizing isoflavones supplements indicate a statistically significant decrease inhot flashes compared to the control treatment (Upmalis et al., 2000; Scambia et al., 2000;Han et al., 2002; Faure et al., 2002). The isoflavones concentrations in these studiesranged from 50 mg to 100 mg per day and resulted in an absolute percent change in hotflash frequency versus the treatment group ranging from 19.5% to 35.8%. The authorsconclude: “Although conclusions based on this analysis should be considered tentative,the available data justify the recommendations that patients with frequent hot flushesconsider trying soyfoods or isoflavones supplements for the alleviation of theirsymptoms” (Messina and Hughes, 2003). In a recent position statement, the NorthAmerican Menopause Society recommends that, for relief for mild vasomotor symptoms,women should first consider lifestyle changes, either alone or combined with anonprescription remedy, such as dietary isoflavones, black cohosh, or vitamin E.

Soy and Brain Function: Fact vs. Fiction

“Soy shrinks the brain” is the proclamation seen on many “anti-soy” websites. Such sitesstate that this sobering soybean revelation is “for real” and not “science fiction.”

The reality is that concerns about soy consumption and brain dysfunction are basedalmost exclusively on a single study from the Honolulu-Asia Aging Study published byDr. Lon White and his colleagues in 2000 (White et al, 2000). The study had manyconfounding factors and limitations as discussed below.

This prospective, epidemiologic (population-based) study involved 3,734 Japanese-American men living in Hawaii who were tested with the Cognitive Abilities ScreeningInstrument (CASI) during a 1991-1994 examination. During that time, the subjects wereasked about tofu consumption at two different time points during midlife: 1965-1967 and1971-1974. The investigators found that poorer cognitive test performance in later lifewas weakly associated with a more oriental midlife diet. More specifically, men whoconsumed tofu approximately 2-4 times per week had odds ratios (OR) in the range 1.6 to2.85 when compared with those eating less than 2 servings/week during the 30 yearfollow-up period. For markers of cognitive impairment (i.e., poor cognitive test scores),the OR was 1.62 (CI 1.06-2.46); for low brain weight the OR was 2.08 (CI 0.97-11.45)and for ventricular enlargement the OR was 2.85 (CI 0.73-11.16).

It must be noted that epidemiologic studies such as this one cannot be used to establishcause and affect relationships; they can only identify associations. This study relies onthe accuracy of tofu intake data and the results may have been influenced by inaccuraciesresulting from the imprecise nature of the methodology employed as discussed in greaterdetail below. The realities of the study conducted by White and colleagues are these:

  • 1.Tofu consumption may be an indicator, not a cause, of other risk factorsassociated with dementia. There are many environmental and genetic factors involved inthe development of dementia. It is possible that other foods or lifestyle factors linkedwith high tofu consumption, not the consumption of tofu per se, were responsible for therelationship found.
  • 2. East Asian countries have a lower incidence of dementia and tend to have a lowerincidence of Alzheimer's disease than do European countries. Therefore, it would beinconsistent to conclude that high tofu consumption (which is indicative of an Asianlifestyle) increases the risk of dementia if rates of dementia increased as a Westernlifestyle is adopted.
  • 3. An article published by Dr. White in the Journal of the American MedicalAssociation in 1996 (White et al., 1996) found that dementia was probably moreprevalent among those men who did not return for cognitive exams compared to thosewho participated. The absence of data from these men could have greatly skewed thestudy results, particularly if they were low tofu consumers. Inclusion of data from thesemen in the study might have shown that dementia levels do not change with the amount of tofu consumed. In fact, in a letter to the editors of the Journal of the AmericanCollege of Nutrition (Guo et al., 2000), it was pointed out that handling of missing datafor 596 individuals in the 2000 study by White could have led to an overestimation of theodds ratios. More specifically, there were 596 individuals who were nonrespondants tothe 1971-1974 interview and hence provide no information for the second measurementof tofu consumption. White and his colleagues assigned missing values for all 596individuals and treated them as reporting having one serving of tofu per week. Thiscould have resulted in a biased overestimation of odds ratios.
  • 4. Many foods in addition to tofu comprise a traditional Asian diet. Dr. White'sstudy looked at only 26 of these foods. In addition, the form documenting dietary intakein 1965 may not have been the best tool to collect these data. The revision of the tool forthe 1971 collection changed the way tofu intake was recorded. Tofu intake may not havebeen accurately measured when these two data sets were combined for Dr. White'sanalysis.
  • 5. Finally, and most importantly, this is only a single study. No public healthrecommendations should even be hinted at, let alone declared, with such a paucity ofdata. More recent research supports the hypothesis that soy isoflavones may have abeneficial, not a harmful, effect on cognitive function. Animal studies have demonstratedthat following dietary administration, soy isoflavones enter the brain in sufficientquantities to activate estrogen receptor B, a newly discovered estrogen receptor to whichisoflavones appear to preferentially bind (Enmark and Gustafsson, 1999). Moreover,administration of soy isoflavones has been shown to improve cognitive function infemale rats (Pan et al., 2000; Lephart et al., 2002).

According to a recent comprehensive report on Phytoestrogens and Health prepared byFood Standards Authority Committee on Toxicity (2003), the report by White et al., 2000did not provide sufficient evidence to confirm the association between high levels ofconsumption of soy-based foods and decreased cognitive function in a group of Japanese-American men and women as the report “lacked sufficient detail and the associationsmay have resulted from inaccuracies in the methods employed.”

More importantly, three clinical intervention studies (one in young men and women andtwo in postmenopausal women have found that a high soy diet (100 mg isoflavones perday) or the use of isoflavone supplements (60 70 mg/day) favorably affected severalaspects of memory and cognition.

In the first study by File et al. (2001), the effects of soy on cognitive function wereassessed in a 10-week placebo-controlled intervention trial of student volunteers aged 22-30 years (15 males and 12 females) who were matched for age, IQ, measures of anxietyand depression and caffeine intake. Subjects consumed calorically equivalent dietscontaining 0.5 or 100 mg total isoflavones per day. Tests of cognitive function wereassessed relative to a pre-study baseline and compared with the placebo group. The testsincluded measures of attention, short-and long-term memory and mood. Subjects in thehigh isoflavone group showed small but statistically significant improvements in tests of short and long term memory (p < 0.05), mental flexibility (p < 0.05) and were treated asmore restrained in a self-assessment of mood (p < 0.05).

A second study by Kris-Silverstein et al. (2001) involving 56 women (ages 55-74)randomly allocated to placebo or soy isoflavones (110 mg/day) for 6 months found thatthose receiving soy showed significantly greater improvement in category fluency, storyrecall, and task planning, suggesting that isoflavone supplementation has a favorableeffect on cognitive function, particularly verbal memory, in postmenopausal women.

In a third study by Duffy et al., (2003), 33 postmenopausal women (50-65 years) notreceiving convention hormone replacement therapy (HRT) were randomly allocated in adouble-blind parallel study to receive a soy supplement (60 mg total isoflavoneequivalents/day) or placebo for 12 weeks. They received a battery of cognitive tests andcompleted analogue rating scales of mood and sleepiness before the start of treatment andthen after 12 weeks. Those receiving the isoflavone supplement showed significantlygreater improvements in recall of pictures and in a sustained attention task. Isoflavonesupplementation also significantly improved learning rule reversals and task planning.

In conclusion, does soy in fact “shrink the brain”? To the contrary, soy isoflavoneconsumption appears to favorably affect several aspects of memory and cognition,although the overall data are too limited to be used for the basis of intakerecommendations at this time (Messina and Messina, 2003).


Ames BN, Profet MN. Gold LS. 1990. Nature’s chemicals and synthetic chemicals:comparative toxicology. Proceedings of the National Academy of Sciences. U.S.A.87:7782-7786. 

Balmir F, Stack R, Jeffrey E, et al. 1996. An extract of soy flour influences serumcholesterol and thyroid hormones in rats and hamsters. Journal of Nutrition 126:3046-3053. 

Chang HC, Doerge DR. 2000. Dietary genistein inactivates rat thyroid peroxidase invivo without an apparent hypothyroid effect. Toxicology and Applied Pharmacology168: 224-252. 

Birk Y. 1993. Protease inhibitors of plant origin and role of protease inhibitors in humannutrition. In: Protease Inhibitors as Cancer Chemopreventive Agents. Troll W, KennedyAR (Eds.), pp. 97-106, Plenum, New York. 

Duncan AM, Merz BE, Xy X, et al. 1999a. Soy isoflavones exert modest hormonaleffect in premenopausal women. Journal of Clinical Endocrinology and Metabolism84:192-197. 

Duncan AM, Underhill KEW, Xu X, et al. 1999b. Modest hormonal effects of soyisoflavones in postmenopausal women. Journal of Clinical Endocrinology andMetabolism. 84:3479-3484 

Eggum BO, Beames RM. 1983. The nutritive value of seed proteins. In: Seed Proteins,Biochemistry, Genetics, Nutritive Value. Gottschalk E, Muller HP (Eds.), pp. 499-531,Martinus Nijhof/Junk, Hague. 

Fang N, Yu S, Badger T. 2004. Comprehensive phytochemical profile of soy proteinisolate. Journal of Agriculture and Food Chemistry. 52:4012-4020. 

FAO/WHO. 1990. Protein quality evaluation. FAO/WHO Nutrition Meetings, ReportSeries 51. Food and Agriculture Organization/World Health Organization, Rome. 

Ikeda T, Nishikawa A, Son HY, et al. 2001. Synergistic effects of high-dose soybeanintake with iodine deficiency, but not sulfadimethoxine or Phenobarbital, on rat thyroidproliferation. Japanese Journal of Cancer Research 92:390-395. 

Jayagopal V, Albertazzi P, Kilpatrick ES, et al. 1997. Beneficial effects of soyphytoestrogen intake in postmenopausal women with type 2 diabetes. Diabetes Care25:1709-1714. 

Kennedy AR. 1993. Overview: anticarcinogenic activity of protease inhibitors. In:Protease Inhibitors as Cancer Chemopreventive Agents. Troll W, Kennedy AR (Eds.),pp. 9-64, Plenum Publishing, NY. 

Kennedy AR. 1994. Prevention of carcinogenesis by protease inhibitors. CancerResearch 54:1999S-2005S. 

Kennedy AR. 1995. The evidence for soybean products as cancer preventive agents.Journal of Nutrition 125:733S-743S. 

Leiner IE. 1995. Possible adverse effects of soybean anticarcinogens. Journal ofNutrition 125:744S-750S. 

Liu K. 1999. Chemistry and nutritional value of soybean components. In: Soybeans,Chemistry, Technology and Utilization, p. 68, Aspen Publishers, Inc., Gaithersburg, MD.McCarrison R. 1933. The goitrogenic action of soybean and ground-nut. Indian Journalof Medical Research 21:179. 

McGuiness EE, Morgan RGH, Wormsley KG. 1984. Effects of soybean flour on thepancreas or rats. Environmental Health Perspectives 56:205-212. 

Mitsuma T, Ito Y, Hirooka Y, et al. 1998. The effects of soybean diet on thyroidhormone and thyrotrophin levels in ageing rats. Endocrinology Regulation 32:183-186. 

Persky VW, Turyk ME, Wang L, et al. 2002. Effect of soy protein on endogenoushormones in postmenopausal women. American Journal of Clinical Nutrition 75:145-153. 

Rackis JJ, Gumbmann MR. 1982. The protease inhibitors: physiological properties andnutritional significance. In: Antinutrients and Natural Toxicants in Foods. Ory RL (Ed.),p. 203, Food and Nutrition Press, Westport, CT. 

Son HY, Nishikawa A, Ikeda T, et al. 2001. Lack of effect of soy isoflavone on thyroidhyperplasia in rats receiving an iodine deficient diet. Japanese Journal of CancerResearch 92:103-108. 

Committee on Toxicity. Committee on Toxicity of Chemicals in Food, ConsumerProducts and the Environment. Phytoestrogens and Health. Food Standards Agency,May 2003 ( 

Fritz WA, Coward L, Wang J, Lamartiniere CA. 1998. Dietary genistein, perinatalmammary cancer prevention, bioavailability and toxicity testing in the rat.Carcinogenesis 19: 2151-2158. 

Hilakivi-Clarke L, Cho E, Clarke R. 1999a. Maternal exposure to genistein duringpregnancy increases carcinogen-induced mammary tumorigenesis in female rat offspring.Oncology Reports 6:1089-1095. 

Hilakivi-Clarke L, Onajafe I, Raygada M. et al. 1999b. Prepubertal exposure tozearalenone or genistein reduces mammary tumorigenesis. British Journal of Cancer 80:1682-1688. 

Lamartiniere CA, Zhao XY, Fritz WA. 2000. Genistein: mammary cancerchemoprevention, in vitro mechanisms of action, potential for toxicity and bioavailabilityin rat. Journal of Women’s Cancer 2: 11-19. 

Messina M, Messina V. 2003. Provisional recommended soy protein and isoflavoneintakes for healthy adults. Nutrition Today 38 (May/June):100-109. 

American Academy of Pediatrics. 1998. Soy protein-based formulas: 

Recommendations for use in infant feeding. Pediatrics 101:148-153. 

Cassidy A, Bingham S, Setchell KDR. 1994. Biological effects of a diet of soy proteinrich in isoflavones on the menstrual cycle of premenopausal women. American Journalof Clinical Nutrition 60:333-340. 

Chang, HC, Doerge, DR. 2000. Dietary genistein inactivates rat thyroid peroxidase invivo without an apparent hypothyroid effect. Toxicology and Applied Pharmacology168:244-252. 

Chen A, Rogan WJ. 2004. Isoflavones in soy infant formula: A review of evidence forendocrine and other activity in infants. Annual Review of Nutrition 24:33-54. 

Cody V, Koehrle J, Hesch RD. 1989. Structure-activity relationships of flavonoids asinhibitors of iodothyronine deiodinase. In: Environmental Goitrogenesis, Gaitan, E. (ed),pp. 57-69, CRC Press, Boca Raton, FL. 

Divi RL, Chang HC, Doerge DR. Anti-thyroid isoflavones from soybean. Biochemicaland Pharmacology 54:1087-1096, 1997. 

Essex C. 1996. Phytoestrogens and soy based infant formulas. British Medical Journal313:507-508. 

Filisetti TM, Lajolo FM. Effect of the ingestion of soybean fractions, raw or autoclaved,on the rat thyroid. Arch. Latinoam. Nutr. 1981;31:287-302. 

Fomon SJ, Ziegler EE. 1992. Isolated soy protein in infant feeding. In: New ProteinFoods in Human Health: Nutrition, Prevention. Steinke FH, Waggle DH, Volgarev MN(eds), pp. 75-83, CRC Press, Boca Raton, FL,. 

Freundlich M, Zilleruelo G, Abitbol C et al. 1985. Infant formula as a cause of aluminumtoxicity in neonatal uraemia. Lancet 2:527-529. 

Freundlich M, Zilleruelo G, Strauss J, et al. 1990. More on aluminum toxic effects inchildren with uremia. Journal of Pediatrics. 117: 1007-1009. 

Klein KO. 1998. Isoflavones, soy-based infant formulas, and relevance to endocrinefunction. Nutrition Reviews 56:193-204. 

Litov RE, Sickles VS, Chan GM, et al. 1989. Plasma aluminum measurements in terminfants fed human milk or a soy-based infant formula. Pediatrics 84: 1105-1107. 

Merritt RJ, Jenks BH. 2004. Safety of soy-based infant formulas containing isoflavones:the clinical evidence. Journal of Nutrition 134: 1220S-1224S. 

Messina M. 2001. Soy and Thyroid Function: Studies Show Little Effect. The SoyConnection 9 (2), Spring issue. 

PDR for Nutritional Supplements. 2001. Hendler SS, Rorvik D (eds.). MedicalEconomics Company, Montvale, NJ. 

Setchell KD, Zimmer-Nechemias L, Cai J, et al. 1997. Exposure of infants to phyto-oestrogens from soy-based infant formula. Lancet 5:23-27. 

Spencer JP, Chowrimootoo G, Choudhury R, et al. 1999. The small intestine can bothabsorb and glucuronidate luminal flavonoids. FEBSLetters 458:224-230. 

Strom BL, Schinnar R, Ziegler EE, et al. 2001. Exposure to soy-based formula ininfancy and endocrinological and reproductive outcomes in young adulthood. Journal ofthe American Medical Association 286:807-814. 

Tran TT, Chowanadisai W, Crinella FM et al. 2002. Effect of high dietary manganeseintake of neonatal rats on tissue mineral accumulation, striatal dopamine levels, andneurodevelopmental status. Neurotoxicology 23:635-643. 

Zung A, Reifen R, Kerem Z, Zadik A. 2001. Phytoestrogens: the pediatric perspective.Journal of Pediatric Gastroenterology and Nutrition. 33:112-118. 

Allred CD, Allred KF, Ju YH. 2001. Soy diets containing varying amounts of genisteinstimulate growth of estrogen-dependent (MCF-7) tumors in a dose-dependent manner.Cancer Research 61:5045-5050. 

Atkinson C, Warren R, Bingham SA, Day NE. 1999. Mammographic patterns as apredictive biomarker of breast cancer risk: effect of tamoxifen. Cancer EpidemiologyBiomarkers & Prevention 8:863-866. 

Atkinson C, Bingham SA. 2002. Mammographic breast density as a biomarker ofeffects of isoflavones on the female breast. Breast Cancer Research 4:1-4. 

Barnes S. 1997. The chemopreventive properties of soy isoflavonoids in animal modelsof breast cancer. Breast Cancer Research and Treatment 46:169-79. 

Bouker KB, Hilakivi-Clarke L. 2000. Genistein: does it prevent or promote breastcancer? Environmental Health Perspectives 108:701-708. 

Fournier DB, Erdman JW, Jr, Gordon GB. 1998. Soy, its components, and cancerprevention: a review of the in vitro, animal, and human data. Cancer EpidemiologyBiomarkers & Prevention 7:1055-1065. 

Hakkak R, Korourian S, Shelnutt SR, et al. 2000. Diets containing whey proteins or soyprotein isolate protect against 7,12-dimethylbenz(a)anthracene-induced mammary tumorsin female rats. Cancer Epidemiology Biomarkers & Prevention 9:113-117. 

Hargreaves DR, Potten CS, Harding C, et al. 1999. Two-week dietary soysupplementation has an estrogenic effect on normal premenopausal breast. Journal ofClinical Endocrinology and Metabolism 84:4107-4024. 

Hsieh CY, Santell RC, Haslam SZ, et al. 1998. Estrogenic effects of genistein on thegrowth of estrogen receptor-positive human breast cancer (MCF-7) cells in vitro and invivo. Cancer Research 58:3833-3838. 

Jakes RW, Duffy SW, Ng FC, et al. 2002. Mammographic parenchymal patterns andself-reported soy intake in Singapore Chinese women. Cancer Epidemiology Biomarkers & Prevention 11:608-613. 

Lamartiniere CA, Moore J, Holland M, Barnes S. 1995a. Neonatal genisteinchemoprevents mammary cancer. Proceedings of the Society of Experimental Biologyand Medicine 208:120-123. 

Lamartiniere CA, Moore JB, Brown NM, et al. 1995b. Genistein suppresses mammarycancer in rats. Carcinogenesis 16:2833-2840. 

Lamartiniere CA. 2000. Protection against breast cancer with genistein: a component ofsoy. American Journal of Clinical Nutrition 71:1705S-1707S. 

Lu LJ, Tice JA, Bellino FL. 2001. Phytoestrogens and healthy aging: gaps in 

knowledge. A workshop report. Menopause 8:157-170. 

McMichael-Phillips DF, Harding C, Morton M, 1998. Effects of soy-proteinsupplementation on epithelial proliferation in the histologically normal human breast.American Journal of Clinical Nutrition 68:1431S-1435S. 

Meng L, Maskarinec G, Wilkens L. 1997. Ethnic differences and factors related tobreast cancer survival in Hawaii. International Journal of Epidemiology 26:1151-1158. 

Messina M, Barnes S. 1991. The role of soy products in reducing risk of cancer.Journal of the National Cancer Institute 83: 541-546. 

Messina MJ, Persky V, Setchell KD, Barnes S. 1994 Soy intake and cancer risk: areview of the in vitro and in vivo data. Nutrition and Cancer 21:113-131. 

Messina M, Loprinzi C. 2001. Soy for breast cancer survivors: A critical review of theliterature. Journal of Nutrition 3095S-3108S. 

Messina M, Messina V. 2003. Provisional recommended soy protein and isoflavoneintakes for healthy adults. Nutrition Today 38:100-109. 

Mommers EC, van Diest PJ, Leonhart AM, et al. 1999. Balance of cell proliferation andapoptosis in breast carcinogenesis. Breast Cancer Research and Treatment 58:163-169. 

Murrill WB, Brown NM Zhyang JX, et al. 1996. Prepubertal genistein exposuresuppresses mammary cancer and enhances gland differentiation in rats. Carcinogenesis17:1451-1457. 

Nagata C. 2000. Ecological study of the association between soy product intake andmortality from cancer and heart disease in Japan. International Journal of Epidemiology29:832-836. 

Petrakis NL, Barnes S, King EB, et al. 1996. Stimulatory influence of soy protein isolateon breast secretion in pre-and postmenopausal women. Cancer EpidemiologyBiomarkers & Prevention 5:785-794. 

Prestin-Martin S, Pike MC, Ross RK, Henderson BE. 1993. Epidemiologic evidence forthe increased cell proliferation model of carcinogenesis. Environmental HealthPerspectives 101 (suppl. 5): 137-138. 

Setchell KDR. 1998. Phytoestrogens: the biochemistry, physiology, and implication forhuman health of isoflavones. American Journal of Clinical Nutrition 68: 1333S-1346S. 

Shao ZM, Wu J, Shen ZZ, Barsky SH. 1998. Genistein exerts multiple suppressiveeffects on human breast carcinoma cells. Cancer Research 38:4851-4857. 

Shimizu H, Ross RK, Bernstein L. 1991. Cancers of the prostate and breast amongJapanese and white immigrants in Los Angeles County. British Journal of Cancer63:963-966. 

Shu XO, Jin F, Dai Q, et al. 2001. Soyfood intake during adolescence and subsequentrisk of breast cancer among Chinese women. Cancer Epidemiology Biomarkers & Prevention 10:483-488. 

Albertazzi P, Pansini F, Bonaccorsi G. et al. 1998. The effect of dietary soysupplementation on hot flushes. Obstetrics and Gynecology 91:6-11. 

Adlercreutz H, Hamalainen E, Gorbach S, et al. 1992. Dietary phyto-oestrogens and themenopause in Japan. The Lancet 339:1233. 

Faure ED, Chantre P, Mares P. 2002. Effects of a standardized soy extract on hotflushes: a multicenter, double-blind, randomized, placebo-controlled study. Menopause9:329-334. 

Gold EB, Sternfeld B, Kelsey JL et al. 2000. Relation of demographic and lifestylefactors to symptoms in a multi-racial/ethnic population of women 40-55 years of age.American Journal of Epidemiology 152:463-473. 

Han KK, Soares JM, Jr, Haidar MA, et al. 2002. Benefits of soy isoflavone therapeuticregimen on menopausal symptoms. Obstetrics and Gynecology 99:389-394. 

Herrington DM, Vittinghoff E, Lin F, et al. 2002. Statin therapy, cardiovascular events,and total mortality in the Heart and Estrogen/Progestin Replacement Study (HERS).Circulation 105:2962-2667. 

Manson JE, et al. 2003. Estrogen plus progestin and the risk of coronary heartdisease. New England Journal of Medicine, 349: 523-534. 

Rossouw et al, 2002. Risks and benefits of estrogen plus progestin in healthypostmenopausal women: principal results from the Women’s Health Initiativerandomized controlled trial. Journal of the American Medical Association 288:321-333. 

Lock M. 1991. Contested meaning of the menopause. The Lancet 337:1270-1272. 

Love SM. 2003. What does it feel like? In: Dr. Susan Love’s Menopause & HormoneBook, pp.46-72, Three Rivers Press, New York. 

Messina MJ. 2003. Potential public health implications of the hypocholesterolemiceffects of soy protein. Nutrition 19:280-281. 

Messina M, Hughes C. 2003. Efficacy of soyfoods and soybean isoflavone supplementsfor alleviating menopausal symptoms is positively related to initial hot flush frequency.Journal of Medicinal Food 6:1-11. 

North American Menopause Society. 2003. Menopause guidebook. The NorthAmerican Menopause Society, Cleveland, OH. 

North American Menopause Society. 2004. Treatment of menopause-associatedvasomotor symptoms: position statement of The North American Menopause Society.Menopause. 11:11-33. 

Scambia G, Mango D, Signorile PG et al. 2000. Clinical effects of a standardized soyextract in postmenopausal women: a pilot study. Menopause 7:105-111. 

Shumaker SA, et al. 2003. Estrogen plus progestin and the incidence of dementiaand mild cognitive impairment in postmenopausal women. The Women's HealthInitiative memory study: A randomized controlled trial. Journal of the AmericanMedical Association 289: 2651-2662. 

Washburn S, Burke GL, Morgan T, Anthony M. 1999. Effect of soy proteinsupplementation on serum lipoproteins, blood pressure, and menopausal symptoms inperimenopausal women. Menopause 6:7-13. 

Upmalis DH, Lobo R, Bradley L, et al. 2000. Vasomotor symptom relief by soyisoflavone extract tablets in postmenopausal women: a multicenter, double-blind,randomized, placebo-controlled study. Menopause 7:236-242. 

Wassertheir-Smoller S. 2003. Effect of estrogen plus progestin on stroke inpostmentopausal women. The Women's Health Initiative: A randomized trial.Journal of the American Medical Association 289: 2673-2684. 

Wuttke W, Jarry H, Becker T. 2003. Phytoestrogens: endocrine disrupters orreplacement for hormone replacement therapy. Maturitas 44 (suppl.):S9-S20. 

Committee on toxicity. 2003. Committee on toxicity of chemicals in food, consumerproducts and the environment. Phytoestrogens and Health. The Food Standards Agency,London. 

Enmark E, Gustafsson JA. 1999. Oestrogen receptors—an overview. Journal ofInternal Medicine 246:133-138. 

File SE, Duffy R, Wiseman H. 2001. Improved memory and frontal lobe function after 3months’ treatment with soy supplements. European Journal of Psychopharmacology157: 430-436. 

Duffy R, Wiseman H, File SE. 2003. Improved cognitive function in post-menopausalwomen after 12 weeks of consumption of a soya extract containing isoflavones.Pharmacology Biochemistry and Behavior 75:721-729. 

Guo G, Wilkens LR, Maskarinec G, Murphy S. 2000. Examining associations of brainaging with midlife tofu consumption. Journal of the American College of Nutrition19:467 (letter to the editor). 

Kritz-Silverstain D, Von Muhlen D, Barrett-Connor E, Bressel MAB. 2003. Isoflavonesand cognitive function in older women: the Soy and Postmenopausal Health in Aging(SOPHIA) Study. Menopause: The Journal of the North American Menopause Society10:196-202. 

Lephart ED, West TW, Weber KS, et al. 2002. Neurobehavioral effects of dietary soyphytoestrogens. Neurotoxicology and Teratology 24: 5-16. 

Messina M, Messina V. 2003. Provisional recommended soy protein and isoflavoneintakes for healthy adults. Nutrition Today 38:100-109. 

Pan Y, Anthony M, Watson S, Clarkson TB. 2000. Soy phytoestrogens improve radialarm maze performance in ovariectomized retired breeder rats and no not attenuatebenefits of 17 B-estradiol treatment. Menopause: The Journal of the North AmericanMenopause Society 7: 230-235. 

White LR, Petrovitch H, Ross W, et al. 2000. Brain aging and mid-life tofuconsumption. Journal of the American College of Nutrition 19:242-255. 

White L, Petrovitch H, Ross GW, et al. 1996. Prevalence of dementia in older Japanese-American men in Hawaii: the Honolulu-Asia Aging Study. Journal of the AmericanMedical Association 276:955-960.