Soy Part 2—Research – Vegan Health

wendy williams vegan diet

Soy Part 2—Research

by Jack Norris, RD • Last updated March 2011

There is a great deal of controversy surrounding soy foods, mostly due to their isoflavones which can bind to estrogen receptors and affect thyroid hormone.

There is significant evidence that eating moderate amounts (one to two servings per day) of traditional soy foods, whether fermented or not, can reduce the risk of prostate cancer and can lower LDL cholesterol.

People have been concerned that moderate amounts of soy could increase the risk of breast cancer or be harmful to women with breast cancer, especially if their cancer is estrogen receptor positive. However, the research to date has been quite reassuring, showing mostly benefits for breast cancer prevention.

Everyone who eats soy should make sure they are getting enough iodine. People with hypothyroidism might need their synthetic thyroid hormone dosage adjusted if they start eating more soy due to the possibility that soy might interfere with it. There is some concern that eating soy could push some people with subclinical hypothyroidism into overt hypothyroidism, so limiting soy for such people might be a good idea.

Soy infant formula has been shown to be safe except possibly for infants with congenital hypothyroidism, whose thyroid function should be monitored. Soy formula is not intended for pre-term infants.

While one observational study found that tempeh was linked to better cognition in older people, tofu has been associated with worse cognition. This is most likely due to confounding variables, including tofu being processed with formaldehyde in Indonesia. Many clinical studies have found that soy increases cognition. Unless you are in Indonesia, you do not need to worry about tofu harming cognition.

Some soy meats or foods containing isolated soy protein are processed with hexane and there may be small amounts of hexane residues in the final product. It is not known if this is harmful, but it might be a good idea to use soy foods from companies who do not use hexane in their processing methods (linked to below).

The phytates in soy can lower the absorption of calcium, zinc, iron, and magnesium. However, you do absorb these minerals from soy foods and eating moderate amounts of soy should not cause deficiencies.

At moderate amounts, soy does not cause feminine characteristics in men. At high amounts, as in twelve servings a day or more, a small percentage of men who are particularly sensitive to soy might develop tender, enlarged breast tissue.

Introduction—Why the Fuss?

By far, soy is the most controversial of plant foods, with gluten taking a distant second place.

Much of the controversy is due to some fairly unique components of soy, isoflavones. Isoflavones are also called phytoestrogens or “plant estrogens”, because they can attach to estrogen receptors in cells. The estrogenicity of soy has raised questions of potential benefits, such as for bone health of post-menopausal women, as well as concerns such as for women with estrogen receptor positive breast cancer.

In addition to the isoflavones, soy also contains phytates that can bind minerals and lower their absorption from foods.

Between 1990 and 2010, there were over 10,000 peer-reviewed journal articles on soy (122). A large percentage of these were conducted in animals which can make the results irrelevant to humans because species differ in how they metabolize soy isoflavones and because the amount of isoflavones given to the animals is often much greater than any human would eat. This much research makes soy one of the most researched foods and also increases the chances of finding results that are outliers – studies that by random chance, or the inability to control variables, could show soy to be harmful (or helpful) when it actually is not. The large number of studies allow people who want to make a case against soy to simply highlight a handful when the bulk of research provides a different view. Of course, someone could make a case in favor of soy in the same way. It is, therefore, important to perform a comprehensive review of the research on any given topic, which I do below for the most important controversies surrounding the potential harm of soy foods.

The Isoflavones

Because some studies look at individual isoflavones, I will give a little background on them. Table 1 shows the breakdown of the typical isoflavone components of soy foods.



In addition to the three main soy isoflavones, there is a fourth isoflavone-related compound, equol, that is produced from daidzin by bacteria that about 25% of Westerners and 50% of Asians and vegetarians have in their digestive tracts (130). Equol has somewhat more estrogenic activity than daidzein and genistein, which has made comparing equol producers the subject of some specific research referenced below.

The isoflavones in soy foods are absorbed almost exactly the same as from supplements (131). After ingesting isoflavones, blood concentrations of genistein and daidzein peak after 5.5 and 7.4 hours respectively; they have a half-life of about 8 hours, meaning 50% will be gone after 8 hours, 75% after 16 hours, and 88% after a day (132). However, if you eat them daily, you will have some in your blood all the time.

There are other isoflavones, the most studied of which are the isoflavones in red clover. These are not the same isoflavones as found in soy.

Soy Servings

In the research discussed here, soy is typically described in grams of protein or milligrams of isoflavones. Less frequently, soy is described in grams of total soy foods. To make things more complicated, sometimes the participants in research are given only soy protein concentrate (about 65% protein), the isolated soy protein (about 90% protein), and sometimes only isolated isoflavones.

A rough guide is that one serving of soy equals 1 cup of soymilk, or 1/2 cup of tofu, tempeh, soybeans, or soy meats. This is the rough equivalent of about 8 to 10 grams of soy protein and 25 mg of isoflavones. The more processed soy meats tend to have more protein (but fewer isoflavones per gram of protein).

Some people who write about soy suggest that fermented soy foods are the most healthy, and that isolated soy proteins are the most unhealthy. Fermented soy foods are tempeh, miso, and natto. As will be discussed below, for the most part, the evidence does not suggest that fermented soy foods are healthier than traditional non-fermented soy foods (such as tofu and soymilk). And much of the research on soy showing benefits has been performed on isolated soy protein.

In the United States, whole soybeans are usually eaten in the form of edamame, tempeh, or soy nuts. Many, but not all, soy meats, as well as texturized soy protein, are made from soy protein concentrate or isolated soy protein. Turtle Island, the makers of Tofurky, use pressed tofu to create their soy meats.

Benefits of Soy

I do not attempt to comprehensively review the benefits of soy in this article. Because they should also be considered, I will briefly cover them.

Lowering Cholesterol

A 2006 review of six meta-analyses on soy and cholesterol levels concluded, “In summary, the systematic reviews suggest that the effect of [a diet containing isolated soy protein with] around 90 mg/day of isoflavones is to reduce LDL cholesterol modestly (by around 5%), without clear effects on triglycerides or HDL cholesterol.” (125)

Prostate Cancer

A 2009 meta-analysis on soy and prostate cancer found that soy, in the highest versus lowest intake categories, was associated with a statistically significant, 26% reduction in prostate cancer risk (0.74, 0.63-0.89) (1). When separately analyzed, studies on non-fermented soy foods yielded a reduced risk of 30% (0.70, 0.56-0.88), and fermented soy foods were not associated with a decreased (or increased) risk of prostate cancer. It should be noted that the highest intake categories in these studies tended to be low, with most being less than one serving per day.

Menopausal Symptoms

A 2009 paper from the 8th International Symposium on the Role of Soy in Health Promotion and Chronic Disease Prevention and Treatment reported that soy isoflavone supplements containing at least 15 mg of genistein per day have been consistently effective at reducing hot flashes (126). If that much genistein is not provided, then there is little benefit, and is the reason why previous reviews, such as a 2006 meta-analysis (5), provided mixed results.

Asian Intakes

Japan and China have lower rates of heart disease and many cancers than do Westerners. People have suggested this could be due to the soy in their diet. In response, soy opponents have often argued that the traditional Asian intake of soy is much lower than is commonly thought. Typical Asian intakes of soy vary among countries and areas. In Japan and Shanghai, China, average intakes are about 1.5 servings per day, but many people consume an average of two or more servings per day. About half the soy eaten in Asia is not fermented.

Below is a review of soy intakes in various Asian countries as described in the 2006 paper by Messina et al, Estimated Asian adult soy protein and isoflavone intakes (116).

Food disappearance data from the Food and Agricultural Organization (FAO) show that daily per capita soy protein intake for Japan between 1961 and 2002 remained constant, and ranged from a high of 9.7 g/day in 1961 to a low of 8.4 g/day in 1977.

The Japanese National Nutrition Survey (NNS) showed that soy intake remained fairly constant after 1960 at about 65 g/day. However, 65 g/day provides approximately 6.5 g of soy protein, which is about 25% lower than the FAO estimate. The total soy food intake of Japanese adults aged 60–69 years is 91.7 g/day, approximately 50% higher than the mean intake of the overall Japanese population.

In recently conducted surveys of individuals in Japan that reported soy protein intake, daily adult female soy protein intake ranged from 6.0 to 10.5 g/day; the male range was 8.0 to 11.3 g/day. Average isoflavone intake ranged from 22 to 54 mg per day among food frequency surveys, and two studies from Japan suggest that approximately 5% of adults consume as much as 100 mg of isoflavones per day. Three studies show that the upper range (rather than the average) of soy protein intake is 16 to 18 g/day.

One study reported that isoflavone intake among subjects from a rural village was similar to that reported for urban Japanese.

Women from the Chinese city of Shanghai have soy intakes about the same as Japan – about 9-10 g/day of soy protein. Isoflavone intake ranged from 33 to 41 mg/day in three studies.

Studies from Hong Kong reported adult female daily soy protein intakes to be 4.9 g/day (2000) and 7.9 g/day (2003). Isoflavone intake ranged from 6 to 30 mg/day in four studies. One study found the upper quartile of women to be consuming 19 g of soy protein per day.

There is evidence to suggest that people in other areas of China may have considerably lower soy intakes, since the national average for total soy food intake in China for 1990 to 1998 was only 17.8 g/day, which represents an isoflavone intake of ≤ 10 mg/day.

Among Chinese women one study found that soymilk, tofu, and processed soy products other than tofu accounted for 81% of total soy protein intake consumed.

Two studies from Singapore, from 2000 and 2002, show relatively low (≤ 5.1 g/day) soy protein intakes. One of those studies showed isoflavone intake to be 16 mg/day; the other didn’t report isoflavone intake.

A 2000 study found average daily isoflavone intake among Korean women aged 35 to 60 years to be 24.4 ± 25.1 mg/day. Intake among pre- and post-menopausal women was 21.8 and 30.1 mg/day, respectively. According to the Korean National Household Survey, the mean daily isoflavone intake (15.1 mg) of individuals living in large cities was almost identical to the isoflavone intake (15.2 mg) of individuals living in rural areas.


In both Japan and China, non-fermented foods provide approximately half of the total soy intake. In Shanghai, nearly all soy is non-fermented.

Breast Cancer

Summary: The vast majority of the evidence is that soy is either neutral or protective against breast cancer, including for women previously diagnosed with estrogen receptor positive breast cancer (tumors stimulated by estrogen contact). This evidence is mostly limited to amounts of two servings per day or less.

Breast Cancer—Reviews

In their article, Soy and Cancer Risk: Our Expert’s Advice, the American Cancer Society says:

So far, the evidence does not point to any dangers from eating soy in people, and the health benefits appear to outweigh any potential risk. In fact, there is growing evidence that eating traditional soy foods such as tofu, tempeh, edamame, miso, and soymilk may lower the risk of breast cancer, especially among Asian women. Soy foods are excellent sources of protein, especially when they replace other, less healthy foods such as animal fats and red or processed meats. Soy foods have been linked to lower rates of heart disease and may even help lower cholesterol.

In their 2010 review, Hilakivi-Clarke et al. sum up the evidence on soy and breast cancer (40):

Results reviewed here suggest that women consuming moderate amounts of soy throughout their life have lower breast cancer risk than women who do not consume soy; however, this protective effect may originate from soy intake early in life. We also review the literature regarding potential risks genistein poses for breast cancer survivors. Findings obtained in 2 recent human studies show that a moderate consumption of diet containing this isoflavone does not increase the risk of breast cancer recurrence in Western women, and Asian breast cancer survivors exhibit better prognosis if they continue consuming a soy diet.

Breast Cancer—Retrospective Studies

In retrospective studies, past diets (provided by memory) of subjects with breast cancer are compared to those without breast cancer. Retrospective studies are less expensive and faster than following people prospectively (forward in time) but are considered less reliable due to inaccuracies in recalling past diet. They generally provide ideas of what connections to investigate prospectively. Nonetheless, the findings from the retrospective studies have shown soy to be either neutral or beneficial in protecting against breast cancer.

California-Hawaii Asian-American Study (2009)

This study looked at childhood soy exposure (18). Women in the top one-third of soy intake during childhood, adolescence, and adulthood had a lower risk of breast cancer than those in the lowest one-third, with the strongest, most consistent effect being for childhood intake. That authors suggest that, “Soy may be a hormonally related, early-life exposure that influences breast cancer incidence.”

Aichi Cancer Center Research Institute, Japan (2008)

This 2008 case control study from Japan found that soy consumption was linked with a significantly reduced risk of estrogen positive breast cancer (8). The risk for the top one-third was .74 (.58 – 0.94).

Nishio et al (2007) sums up the results of the previous restrospective studies (17):

“…seven studies showed no associations [between soy and breast cancer], one study showed protective effects, six studies showed protective associations only in premenopausal women, and one study found [protective] associations only in postmenopausal women.”

Breast Cancer Observational Studies of Women Initially without Cancer

In typical observational studies, you start with people who are apparently healthy and follow them prospectively, without any clinical intervention. The observational studies on soy and breast cancer can be divided into two groups—those done on populations with very low intakes and those done on populations with higher intakes.

The studies done on populations with very low intakes do not provide much evidence. Since 2001 there have been six observational studies on such populations. One of those studies, EPIC-Norfolk (2004), did find an association between soy intake and breast cancer, but the intake of soy was so small as to make the finding likely due to random chance.

EPIC-Norfolk, Phytoestrogens (2010)

The European Prospective Investigation into Cancer and Nutrition (EPIC), Norfolk chapter, measured phytoestrogen content of all foods (not only soy) and compared intakes to breast cancer rates (19). They found no association between phytoestrogens and breast cancer.

French Cohort (2006)

Women were followed from 1990-1991 through July 2002. No relationship between isoflavone intake and breast cancer risk was found. Isoflavone intake was extremely low with the highest group in the range of .03 to .1 mg/day (25).

EPIC-Norfolk (2004)

EPIC-Norfolk performed a nested case-control study measuring isoflavone intake and risk of breast cancer. Women were recruited from 1993-1997 and followed through July 2001 (20). Isoflavone intake was very low, with an average intake of .4 mg/day. Urine and serum isoflavone levels (taken at the beginning of the study) were compared to breast cancer rates. Exposure to equol and daidzein was associated with a statistically significant increased risk of breast cancer. This is the only prospective finding in which soy was associated with an increased risk of breast cancer; because the intakes of soy were so low, the finding is most likely an anomaly.

The Netherlands (2004)

This study found no association between soy intake and breast cancer. Intake of soy was very small, only 4.9 g of soy foods per day in the highest one-fourth (29).

California Teachers Study (2002)

This study, that took place in Los Angeles, found no association between soy intake and breast cancer, but it only had two years of follow-up and isoflavone intake was 2 mg/day at the 80th percentile – just a fraction of a serving per day (28).

The Netherlands (2001)

This nested case-control study found that urinary genistein levels collected 1 to 9 years before post-menopausal breast cancer was diagnosed was not associated with risk (27).

Observational Studies with Higher Soy Intakes

Of the studies done on populations with higher soy intakes (about one to two servings per day as a typical upper intake amount), the Singapore Chinese Health Study (21), the Shanghai Women’s Study (22, 23), and the Japan Public Health Center (26) study all found that higher intakes of soy were associated with a reduced risk of cancer. The Japan Collaborative Cohort Study (17) and the Japan Life Span Study (30) found no association. European Prospective Investigation into Cancer-Oxford (24) which contained a large number of vegetarians, also found no association, but this could be because exposure to soy might be more protective when breast tissue is developing during adolescence, while Western vegetarians often come to the diet as adults.

Singapore Chinese Health Study (2010)

This report of the Singapore Chinese Health Study was a bit different (than the 2008 report below) in that it examined dietary patterns, and not just soy intake (21). There were two diet patterns which had previously been identified as common among Asian women: “vegetable-fruit-soy” and “meat–dim sum.” The women were recruited from 1993 to 1998, had a median age of 55 years, and were followed through 2005. Women who ate closer to the vegetable-fruit-soy pattern had a lower risk of breast cancer, but this result was limited to women who were post-menopausal at the beginning of the study. There were a number of non-statistically significant trends towards less risk, including for soy and isoflavone intake, fruit intake, vegetable intake, and vegetable-fruit-soy intake for estrogen receptor positive breast cancer.

Shanghai Women’s Study (2009)

The Shanghai Women’s Study (SWS) found that among pre-menopausal women, higher soy intake was associated with a lower risk of breast cancer (23). A soy protein intake of 16 g/day was associated with an almost 60% reduced risk when compared to the lowest intake of soy protein of 3.5 g/day or less. SWS also surveyed soy intake during adolescence (based on the participants’ memory) and found it to be associated with a lower risk of pre-menopausal breast cancer.

Singapore Chinese Health Study (2008)

Women were recruited from 1993 to 1998 and followed through the end of 2005 (22). The women were divided into those eating less than 10.6 mg of isoflavones per 1,000 calories and those eating 10.6 mg or more. Women in the higher intake group had a lower risk of breast cancer (.82, .70-.97). When divided by pre- and post-menopausal, the post-menopausal women (only) had a lower risk (.74, .61 – .90). Post-menopausal women who ate more soy had a lower risk of estrogen receptor positive cancer (.67, .49–.91). The authors noted that “There are seven common soy products in the Singapore Chinese diet and all are nonfermented.”

EPIC-Oxford (2008)

EPIC-Oxford followed 37,600 British women, 31% of whom were vegetarian, for 7.4 years (24). There was no difference in risk of breast cancer between high isoflavone intake (average of 31.6 mg/day) and low isoflavone intake (.2 mg/day). No significant associations were observed when separated into pre- and post-menopausal groups.

Japan Collaborative Cohort Study (2007)

The Japan Collaborative Cohort Study found no correlation between the consumption of tofu, boiled beans, or miso soup and the risk of breast cancer (17).

Japan Public Health Center (2003)

The Japan Public Health Center-Based Prospective Study on Cancer and Cardiovascular Diseases found that people in the highest one-fourth of genistein intake of 24 mg/day had only half the risk of breast cancer compared to those in the lowest genistein intake of 7 mg/day (26). Those in the highest soyfoods intake category (“almost every day”) didn’t have a lower risk compared to those in the lowest (“less than 2 times per week”). Upon further analysis, the finding for isoflavones being protective was limited to post-menopausal women, for whom there was a 68% lower risk in the highest isoflavone intake category.

Life Span Study, Japan (1999)

The Life Span Study cohort in Hiroshima and Nagasaki found no association between tofu and miso with breast cancer risk. The highest soy intake category was 5 or more servings per week (30).

Breast Cancer Survival and Recurrence

The Women’s Healthy Eating and Living (2011) study is a randomized controlled trial of a high fruit-vegetable-fiber and low fat dietary intervention in early stage breast cancer survivors (121). It had a median follow-up of 7.3 years from the time of enrollment. Soy intake was measured post-diagnosis (median 2 years, range: 2 months to 4 years) using a food frequency questionnaire that included specific items for “Meat Substitutes (such as Tofu, Veggie Burgers),” and “Soy Milk”, as well as an opportunity to include other soy foods and supplements.

Isoflavone intake (the marker for soy) was unrelated to the risk of recurrence regardless of hormone receptor status or Tamoxifen use. No significant increased or decreased risk was associated with any specific level of intake. Risk of death tended to be lower as isoflavone intake increased (p for trend=0.02). Women at the highest levels of isoflavone intake (>16.3 mg/day isoflavones; equivalent to at least 1⁄2 cup soymilk or 2 oz tofu) had a non-significant 54% reduction in risk of death compared to the lowest one-fifth of soy intake.

The authors state:

Our study is the third epidemiological study to report no adverse effects of soy foods on breast cancer prognosis. These studies, taken together, which vary in ethnic composition (two from the US and one from China) and by level and type of soy consumption, provide the necessary epidemiological evidence that clinicians no longer need to advise against soy consumption for women diagnosed with breast cancer.

Of the other studies, one from the Cancer Hospital of Harbin Medical University, China (2010) (31), the Shanghai Breast Cancer Survival Study (2009) (11), and the Long Island Breast Cancer Study (2007) (32) found soy to be associated with a lower risk of breast cancer or death in some subgroups. Neither those studies nor the Shanghai Breast Cancer Study (2005) (33) or the Life After Cancer Epidemiology Study from Northern California and Utah (2009) (34) found any statistically significant increase in breast cancer, including among women with estrogen positive breast cancer (11, 33, 34).

Breast Tissue and Nipple Aspirate

Non-lactating nipple aspirate might be a risk for breast cancer, although the data is mixed, and it is likely that what is more important is whether certain types of cells are found in the aspirate (129).

There have been three studies measuring the effects of soy on nipple aspirate. One was conducted in women diagnosed with breast cancer and it found increased aspirate after two weeks in the high soy diet, although there was no increase in breast cell proliferation (35). In two longer term studies of six months on women without a diagnosis of breast cancer, one study found increased nipple aspirate in women while on a higher soy diet of 45 mg of isoflavones per day (there was no placebo group) (37). The other found no increase in nipple aspirate on the high soy diet (2 servings per day) when compared to the low soy diet (

Almost everyone who has a reliable source of iodine can safely eat soy without it causing thyroid problems. Although most studies that have measured thyroid function and soy intake have found no problems, a 2011 study (78) of people with subclinical hypothyroidism found an increased rate of progression to overt hypothyroidism. For such people, it might be wise to limit soy. People with overt hypothyroidism who substantially change their soy intake might need to talk to their doctor about adjusting their synthetic thyroid medication.


The thyroid gland produces hormones, triiodothyronine (T3) and thyroxine (T4), that regulate metabolism. Thyroid stimulating hormone (TSH), made by the anterior pituitary gland, increases in response to the body’s need for more T3 or T4.

Isoflavones in soy can inhibit thyroid peroxidase (TPO) (77), an enzyme involved in the synthesis of the thyroid hormones. This mostly happens when someone is deficient in iodine. If left unchecked, this can lead to hypothyroidism and even an enlarged thyroid gland, known as a goiter.

In their 2006 review article, Messina and Redmond write (77):

The preponderance of evidence from clinical trials involving healthy adult men and women indicates that neither soy protein nor isoflavones adversely affect thyroid function. As noted the adverse effects reported by one Japanese study [Ishizuki Thyroid Clinic, (67)] are biologically implausible and contrast with the results of 13 other trials. Thus, despite their ability to [inhibit thyroid peroxidase] in vitro and in vivo in rodents, isoflavones do not appear to cause thyroid hormone abnormalities in euthyroid individuals [people with genetically normal thyroid function]. Nevertheless, research aimed at determining whether further depression of thyroid function in individuals with pre-existing subclinical hypothyroidism occurs in response to soy intake is warranted.

Messina and Redmond also point out that soy might interfere with the absorption of synthetic thyroid hormone.

In their 2002 review article, Doerge and Sheehan write (76):

Iodine deficiency is an emerging concern in elderly Americans. Consumption of iodized salt, the primary source of dietary iodine, may decrease with the desire or need to reduce the possible hypertensive effects of high salt intake. The data presented here suggest that elderly women need to be aware of, and monitored for, possible thyroid problems resulting from consumption of soy products. Those post-menopausal women who consume large amounts of soy products may be at higher risk.

Primary Studies

Hull Royal Infirmary, United Kingdom (2011)

This study was done on people who were found to have subclinical hypothyroidism (defined as a TSH value between 5 and 15 mU/liter; normal range is 0.5–4.7 mU/liter) (78).

It was a randomized, double-blinded, crossover study of 60 patients (8 men, 52 women), ages 44–70. The two study periods lasted eight weeks each and consisted of 30 g of soy protein that provided either 2 mg of isoflavones or 16 mg of isoflavones with an eight week washout period in between. An isoflavone intake of 2 mg per day is typical of Western populations, while 16 mg is typical of vegetarian intakes. The subjects had normal iodine levels ( >100 µg/liter).

Six patients, all women, developed overt hypothyroidism after the 16 mg isoflavone phase, while none did after the 2 mg phase. The overt hypothyroidism appeared permanent, with no reversal on isoflavone withdrawal, although to determine this fully they should have discontinued thyroid hormone for a period, which they did not do. The underlying cause is unclear, although acceleration of an underlying autoimmune process is a possibility.

The authors state:

In a prospective study looking into the spontaneous course of hypothyro >

The authors conclude that female vegetarian patients with subclinical hypothyroidism may need careful monitoring.

The good news was that the high isoflavone phase produced a significant reduction of blood pressure, c-reactive protein, and insulin resistance. It should also be noted that this is just one study which is rarely enough to be considered conclusive.

Secondary Studies

The studies below, while not primarily looking at thyroid function, did measure it while studying other effects of soy.

Studies showing no effects on thyroid hormone were:

  • 54 mg/day genistein in post-menopausal women for 3 years (13)
  • 76 mg/day isoflavones in post-menopausal women for 2 years (75)
  • 10 mg/day of equol in post-menopausal women for 1 year (63)
  • 90 mg/day isoflavones on post-menopausal women for 6 months (60)
  • 65 mg/day isoflavones in men and women for 3 months (69)
  • 48 mg/day isoflavones on children for 2 months (66)
  • 62 mg isoflavones on men for 2 months (68)
  • 54 mg/day isoflavones on post-menopausal women for 7 weeks (71)

A study on soy’s effects on protein synthesis by the liver showed no effect on thyroid binding globulin (TBG) after 118 mg/day isoflavones for 3 months in post-menopausal women (72).

The following studies showed minor changes in thyroid hormone but were not believed to be of physiological significance:

  • 90 mg/day isoflavones in post-menopausal women for 6 months (73)
  • 132 mg/day isoflavones in post-menopausal women for 3 months (70)
  • 128 mg/day isoflavones in pre- and post-menopausal women for 3 months (61)
  • 40 mg/day isoflavones in pre-menopausal women for 1 month (74)
  • 140 g/day soybeans in female students for 1 week (65)

One study of 50 g/day soy protein in men for 4 weeks found that T4 changed from baseline, but did not differ from the placebo group (64).

Finally, one study found that 30 g/day of pickled soybeans didn’t change thyroid hormone levels but caused a goiter in half the subjects.

Infant Formula

Summary: The American Academy of Pediatrics and the National Toxicology Program considers soy formula safe for term infants. There has been one cohort study looking at the general health of adults who were fed soy formula as infants. It found no reason to be concerned about thyroid or reproductive function (95). The Beginnings Study is an ongoing study examining the effects of formula on development. It is in its early stages with findings from children only a year old, but to date no negative effects of soy have been found on growth, sex organs, or neurological development compared to children on cow’s milk formula. It is best to choose a soy formula with DHA. Soy-formula is not intended for pre-term infants.

Excerpts from Setchell et al’s 1997 review paper on soy infant formulas sums up the controversy (79):

When bodyweight and milk intake are taken into account, an infant exclusively fed soy-based formulas would be exposed to a fairly constant dose [of isoflavones] throughout early life. This dose represents a 6–11 fold higher level of intake of isoflavones than the bodyweight-adjusted intake (0.7 mg/kg per day) found to cause significant modifications to the hormonal regulation of the menstrual cycle of western women…. These values are also much higher than plasma isoflavone concentrations of Japanese adults, which range from 40 to 240 ng/mL.

Seven years later in 2004, Merritt and Jenks wrote (82):

Largely as a result of research in animal models, concerns have been voiced regarding isoflavones in soy infant formulas in relation to nutritional adequacy, sexual development, neurobehavioral development, immune function, and thyroid disease. We discuss the available clinical evidence regarding each of these issues. Available evidence from adult human and infant populations indicates that dietary isoflavones in soy infant formulas do not adversely affect human growth, development, or reproduction.

While the American Academy of Pediatrics Committee on Nutrition (2008) recommends not using soy formula unless cow’s milk formula is contraindicated or the parents are vegan, they say that there is no reason to think that it harms development in term infants (84):

In summary, although studied by numerous investigators in various species, there is no conclusive evidence from animal, adult human, or infant populations that dietary soy isoflavones may adversely affect human development, reproduction, or endocrine function… [A]n increased incidence of feminization in male infants or an increased incidence of hypospadias in high soy-consuming populations have not been observed.

Regarding infants with thyroid problems:

Consumption of soy products by infants with congenital hypothyroidism complicates their management, as evidenced by a prolonged increase in thyroid-stimulating hormone when compared with infants not fed soy formula; the authors of 2 studies suggested closer monitoring and a possible need for an increased dose of levothyroxine.

The American Academy of Pediatrics does not recommend soy infant formulas for pre-term infants due to studies showing poor bone growth in pre-term infants on soy-formula compared to cow’s milk formula designed for pre-term infants. There is also a concern that aluminum levels in soy formula could be a problem for infants with undeveloped kidneys. They say, “Term infants with normal renal function do not seem to be at substantial risk of developing aluminum toxicity from soy protein-based formulas.”

The National Toxicology Program’s Center for the Evaluation of Risks to Human Reproduction convened an expert panel on December 16-18, 2009 to evaluate soy infant formula (85). They concluded that there is minimal concern for adverse developmental effects in infants fed soy formula.

Other studies include:

  • Three case studies of infants with goiter in 1960, before soy formula was fortified with
    iodine (80).
  • A 1990 retrospective case-control study showing children diagnosed with autoimmune thyroid
    disease were more likely than controls to have received soy formula as infants (81).
  • A 2001 study of young men and women who had eaten soy infant formula found
    no reason to be concerned about thyroid or reproductive function, although the women had
    slightly longer duration and greater discomfort of menstrual bleeding (95).
  • A 2008 cross-sectional study showing that female toddlers who received soy
    had more breast bud tissue left (93).
  • A 2008 cross-sectional study found female girls on soy formula had a higher
    maturation index which the authors thought should be examined in future studies (94).
  • A 2009 cross-sectional study found no difference in sex hormone levels in children fed soy
    formula (83).
  • A 2010 study found that women who had received soy formula as infants were more likely to
    have non-cancerous uterine fibroids (92).

Summary: There have been 12 short-term clinical trials looking at the impact of soy on cognition, and all have shown soy to be helpful or neutral. A longitudinal study found tempeh to be associated with improved cognition. Three reports from longitudinal studies have associated tofu with reduced cognition in some groups, but increased cognition in another group, and neutral in others. This is likely due to confounding. Based on the research to date, there should be little concern about eating soy, including tofu, with regards to cognitive decline.

Long-term Studies

In the Honolulu-Asia Aging Study (2000) of older Asian men living in Hawaii, food intake was collected once at baseline (1965–1967) and one more time during 1971–1974. Cognitive function was then measured in 1991–1993, and brain size was measured in those who had died (2).

Midlife tofu consumption by itself (controlling for no other factors) explained 2.3% of the variation in mental cognition scores. A linear regression model that included only age, education and history of a prior stroke explained 27.8% of the variation. After controlling for these three most important factors, midlife tofu consumption remained statistically significant even though it then explained only 0.8% of the variance in test scores. Despite this, the researchers state that, “In this study population, 20% to 25% of the burden of cognitive impairment appears attributable to midlife tofu consumption — an effect size of enormous public health importance…”

Upon autopsy of those who died, smaller brain size was associated with increased tofu consumption.

In an accompanying editorial, Grodstein et al, write (41):

Clearly, these results are interesting, although must be considered preliminary; relatively few subjects consumed tofu at high levels, and the confidence intervals around the estimates of effect presented are wide, indicating the limited precision of these results… Importantly, we do not know if tofu itself was the cause of these numerous indications of accelerated brain aging or if tofu is merely a marker for some other unfavorable exposure. For example, in this population, men with higher tofu intake (and more traditional diets) likely came from poorer immigrant families and perhaps experienced more childhood privation, which may be related to their brain development and subsequent cognitive function.

In 2010, Hogervorst et al, reference this study (42):

People of low socioeconomic status were hypothesized to eat more soy (as an animal protein substitute because it is cheaper). However, low socioeconomic status in itself is associated with a faster cognitive decline and an earlier onset of dementia and could thus have mediated the association between high soy intake and low memory function.

The Kame Project (King County, Washington, 2000) is a prospective study of Japanese Americans aged 65 years or older living in King County, WA (53). Cognitive ability was measured at the beginning and after two years. Tofu consumption was categorized as low (


The short-term clinical trials have been very favorable towards soy and cognitive function. The one long-term study that assessed a non-tofu soy food, tempeh, found a protective effect in the first analysis, and no effect after adjustments in a follow-up.

The main concern appears to be tofu which has recently been prepared with formaldehyde in Indonesia and was a food traditionally eaten by poorer people with less education. The research can be summarized as follows:

  • The Honolulu-Asia Aging Study found tofu consumption during the 1960s and 1970s was associated with cognitive decline in the 1990s (2).
  • An Indonesian study of older people found that tofu was associated with slightly lower memory scores, possibly due to formaldehyde in processing (9). A follow-up report found that tofu was linked with better memory scores among the younger subjects (42).
  • The Kame Project, a study of older Japanese Americans found that at baseline, and only among women taking estrogen replacement therapy, tofu consumption was associated with lower cognitive scores. Testing again two years later, this association was no longer there (53).

It seems safe to assume that tofu does not have an appreciable effect on cognitive function and that these associations are due to confounding variables. A study on non-Asians and tofu consumption (where it cannot be linked to lower socioeconomic status or formaldehyde) could shed more light on this subject.

There is little evidence to determine whether the hexane residues in some processed soy meats are safe over the long term, though hexane processing can be harmful to workers and the environment. It seems prudent to buy most of your processed soy from companies that do not use hexane extraction.

Hexane is sometimes used by the soy industry to separate the oil from the protein in soybeans. The protein is then used for soy meats and other products that contain soy protein extract. This process inevitably leaves some hexane residues in the products.

The Cornucopia Institute is a public interest group that promotes organic agriculture. In November 2010, they released a report Toxic Chemicals: Banned In Organics But Common in “Natural” Food Production. The report makes a persuasive argument that hexane use is dangerous to workers and the environment (the evaluation of which is beyond the scope of this article), and for those reasons alone it seems better to use soy meats processed without hexane. Luckily, there are soy meats available that do not use hexane such as Tofurky, Field Roast, Wildwood, Amy’s Kitchen, and some (though not all) Boca products (link). Foods certified as “USDA Organic” are produced without using hexane extraction.

There is not much evidence that the amount of hexane found in most veggie burgers is harmful. The U.S. Food and Drug Administration has not set an upper limit on how much hexane residue can be in foods, whereas the European Union prohibits hexane residues greater than “30 mg/kg [30 ppm] in the defatted soya products as sold to the final consumer” (118). The Cornucopia Institute found some soy ingredients to have hexane levels as high as 21 ppm, but did not provide details on the average hexane levels they found in any final food products.

More info on hexane in soyfoods:

  • Hexane in Soy Food – Berkeley Wellness • May 01, 2012
  • Do Veggie Burgers Contain Hexane? – Shereen Jegtvig, MS • Updated Feb 07, 2014

Mineral Absorption

Eating a few servings of soy per day is unlikely to result in calcium, zinc, iron, or magnesium deficiency.

Table 2. Phytate Content of Selected Foods 97

Table 1. Typical Isoflavone Content of Soy 122
Isoflavone If Fermented % of Total
genistin genistein
daidzin daidzein
glycitin glycitein
Food Range or Average g/100 g
Wheat 0.52 – 1.05
Corn 0.62 – 1.17
Barley 0.85 – 1.18
Oats 0.90 – 1.42
Peas 0.72 – 1.23
Soybean 1.20 – 1.75
Soy isolate 0.82
Lupine bean 0.76 – 1.63
Vital gluten 1.90

Soybeans contain phytic acid, also known as phytate, which can inhibit the absorption of calcium, zinc, iron, and possibly magnesium. Many whole plant foods contain phytate, but soy has more than most. Hidvegi and Lasztity estimated the phytate content per 100 grams of food (97) in Table 2.

While phytic acid has a bad reputation, it has some benefits. In their 2002 review, Minerals and Phytic Acid Interactions: Is It a Real Problem for Human Nutrition?, Lopez et al point out that phytates can prevent lipid peroxidation, iron oxidation of colorectal tissue, and calcium-based kidney stones (98).

Three studies on adults (99, 100, 101) have shown calcium to be absorbed from soy at rates comparable to that of cow’s milk. One study on Chinese boys found calcium from soy milk to be absorbed at 43% while calcium from cow’s milk was absorbed at 64% (109, see Table 3).

A study from Victoria University (2010) found calcium absorption to be the same from fortified soymilk as from cow’s milk in post-menopausal women (101).

A study from Purdue University (2005) found that the calcium in soymilk fortified with calcium carbonate was absorbed at the same rate (21%) as the calcium in cow’s milk (100). The calcium in soymilk fortified with tricalcium phosphate was absorbed at a lower rate of 18%.

In a study from Creighton University (1991), soybeans, grown in different mediums to produce a different phytate content, were used to determine calcium absorption (99). They found that calcium from high-phytate soybeans was absorbed at 31%, while calcium from low-phytate soybeans was absorbed at 41%. Calcium from milk was absorbed at 38%.

Table 3. Mineral Absorption Rates in Chinese Boys 109
Food Calcium
Soy milk 43.5 11.3 6.7
Cow’s milk 64.2 31.2 15.5
Dephytinized soy milk 50.9 20.1 20.6

Soy and Bones

Although a 2010 (3) and 2008 (4) meta-analysis each found that soy isoflavone supplements of 82 and 90 mg/day, respectively, increases bone mineral density in menopausal women, a more recent, clinical trial found little benefit from 80 or 120 mg of soy isoflavone supplements for three years (6).

There is a plethora of evidence that soy does not harm bones; including a cross-sectional study from the Chinese University of Hong Kong (2003) which found that women who ate the most soy (10 g/day of soy protein or more) had greater bone mineral density than those in the lower intake groups (104).

Studies have shown that the zinc in soy is absorbed at a rate of about 10–20%.

A 2004 study from Switzerland showed that removing the phytic acid from a wheat and soy flour increased zinc absorption from 22.8% to 33.6% in healthy adults (108). Copper absorption was unaffected.

Table 2 shows results from a 2003 absorption study on Chinese boys (109). Zinc absorption rates were lower from soymilk than cow’s milk.

A study from University of Gothenburg in Sweden (1987), found that adding milk to a meal with meat sauce reduced zinc absorption from 25.2% to 20.7% (110). Adding soy to a meal with meat sauce decreased zinc absorption to 18-20%. Meals of only soy had zinc absorption rates of 14-21%.

Iron from plant foods is generally not absorbed as well as iron from meat. But a type of iron in soy, ferritin iron, is absorbed at about 30% among people with low iron stores, and this is a high rate of absorption. Ferritin iron makes up a large percentage of the iron found in soybeans; up to 90% (127). Some soy foods have a decent amount of iron, including extra firm tofu (3.35 mg per 1/2 cup), edamame (1.75 mg per 1/2 cup), and soy nuts (1.7 mg per 1/4 cup). There is no reason to worry that moderate amounts of soy might cause iron deficiency.

Also note that adding vitamin C to a meal (in a dose of about 100 mg) has been shown to significantly increase absorption of the iron in plant foods in numerous studies. I found no studies on vitamin C’s effect on iron absorption from soy foods in adults, but two showed that it significantly increased iron absorption from soy formula in infants (114, 115).

I found no studies looking at the effect of soy on magnesium absorption. Phytates have been shown to bind magnesium, so it would be no surprise if the magnesium in soy is absorbed at a lower rate than from other foods. Leafy green vegetables, whole grains, nuts, and seeds are the best sources of magnesium and people eating such foods while adding a few servings of soy per day should not be in danger of magnesium deficiency.


In their 2007 review, Hotz and Gibson say that fermentation can reduce the amount of zinc and iron-binding phytates in soybeans and should, therefore, increase the amount of zinc and iron absorbed from fermented soy foods such as tempeh and miso (111).

The effects of fermenting soy foods on rates of zinc absorption have not been tested in humans, but one study from the University of the Witwatersrand, South Africa (1990) found that the iron in tempeh was absorbed better than from other soy foods, at a rate of 10 – 15%; the iron from tofu was absorbed at a range of about 3 to 8% (113). Again, absorption rates of plant iron vary greatly according to someone’s current iron status and the amount of iron in the food, which can account for a large range of absorption amounts.

Feminizing Characteristics

It requires twelve servings of soy (and probably much more for most men) to have any sort of noticeable feminizing effects. While one epidemiological study raised concerns about soy and sperm quantity (14), two clinical studies have shown no effects of soy on sperm quality or quantity (15, 126).

In 2011, there was a case report of a 19-year old vegan male who was eating a great deal of soy foods – enough to provide 360 mg of isoflavones per day, the equivalent of about 14 servings (10). He also had type 1 diabetes. After eating this way for a year, he developed low free testosterone levels and erectile dysfunction. After ceasing the soy products (and the vegan diet), his symptoms normalized within a year. There was no reason why he needed to stop being vegan rather than just cutting way down on, or eliminating for a period, soy foods.

A 2008 case report described a 60 year old man who developed gynecomastia (the enlargement of the mammary glands in a male) after drinking 3 quarts of soymilk (the equivalent of 12 servings) a day (123). His breasts returned to normal after discontinuing the soymilk.

Gynecomastia – Male Breasts

In contrast to the case report above, in a 2004 study, 20 men with prostate cancer were given either 450 mg (amount found in about 18 servings of soy) or 900 mg (amount found in about 36 servings of soy) of isoflavones for 84 days (124). Two of the men, who had no gynecomastia at baseline, receiving the 900 mg developed mild cases of gynecomastia. One man was taking a drug that likely increased the gynecomastia. Two men had hot flashes probably related to the isoflavones. Given the very large amounts of isoflavones they were taking, this study should allay fears of feminization in men rather than cause any sort of alarm.

Sperm Health

A 2008 epidemiological study found that men attending a fertility clinic had lower sperm concentrations if they ate more soy, but the amount of soy was very small (≥ .3 servings/day) and part of the difference could be explained by a higher ejaculate volume (14). Two more recent clinical trials using 62 mg/day of isoflavones for two months and 480 mg/day of isoflavones for three months found no difference in sperm count, concentration, or motility.

Messina sums up the rest of the literature on soy and feminizing characteristics in men in his 2010 review (122):

In contrast to the results of some rodent studies, findings from a recently published metaanalysis and subsequently published studies show that neither isoflavone supplements nor isoflavone-rich soy affect total or free testosterone (T) levels. Similarly, there is essentially no evidence from the nine identified clinical studies that isoflavone exposure affects circulating estrogen levels in men. Clinical evidence also indicates that isoflavones have no effect on sperm or semen parameters, although only three intervention studies were identified and none were longer than 3 months in duration. Finally, findings from animal studies suggesting that isoflavones increase the risk of erectile dysfunction are not applicable to men, because of differences in isoflavone metabolism between rodents and humans and the excessively high amount of isoflavones to which the animals were exposed. The intervention data indicate that isoflavones do not exert feminizing effects on men at intake levels equal to and even considerably higher than are typical for Asian males.


A 2009 meta-analysis of five retrospective and two prospective studies found that participants with a higher soy intake (roughly one serving per day) had a reduced risk for endometrial cancer and ovarian cancer, when compared with lower soy intakes. Two other studies of three years in duration found that soy isoflavones in the amounts of 70, 80, and 120 mg/day did not have a negative effect on the endometrium. However, one study using 90 mg of isoflavones for five years (equivalent to 3.5 servings of soy per day) did result in cell growth (though non-cancerous) of the endometrium among some subjects. Women at risk for endometrial cancer might use caution in eating more than one serving per day of soy.

Ovarian Function

A 2010 review by Wendy N. Jefferson of the Laboratory of Reproductive and Developmental Toxicology, National Institute of Environmental Health Sciences, summarized the literature on soy and ovarian function (117):

For the most part, the studies conducted to date suggest that a diet containing lower levels of soy, e.g. 1–2 servings of soy/d, as part of a well-balanced diet should not pose harmful effects on the function of the ovary as it relates to ovulation. These levels are similar to that found in a traditional Asian diet (10–25 mg/day isoflavones) and even up to 50 mg/day isoflavones has little impact on serum circulating levels of hormones involved in reproduction. Although the levels of phytoestrogens typically found in soy foods pose minimal risk in the adult female, the female reproductive system is dependent on hormones for proper function and phytoestrogens at very high levels can interfere with this process.

Additionally, a 2009 meta-analysis of five retrospective and two prospective studies found that participants with a higher soy intake (roughly one serving per day) had a reduced risk for endometrial cancer and ovarian cancer, when compared with lower soy intakes (119).

A 2014 cross-sectional study from Adventist Health Study-2 found that a higher intake of isoflavones was associated with an increased rate of never becoming pregnant and of being childless (136). Among the highest soy intake category of > 50 mg/day (average of 79 mg/day or about three servings), compared to those in the lowest intake category (average of 3.7 mg/day), there was a 15% higher rate of being childless (1.15, 1.02-1.29). This finding was adjusted for age, educational level, and marital status.

The authors did not distinguish between involuntary and voluntary childlessness.

This issue warrants further attention, but given that it is a cross-sectional study and the findings were rather weak, it seems unlikely that isoflavones cause childlessness.

Isolated soy protein is rumored by some to have unusually large amounts of MSG (it doesn’t), and MSG, in turn, is rumored to be an excitotoxin (for practical purposes, it’s not) that will destroy brain cells. I cover this topic in some blog posts at (link).


Two studies have examined soy intake and risk of hypospadias. Neither found a statically significant risk with more soy intake. It appears that factors other than soy are the most likely the cause of hypospadias, although more studies are warranted.

Hypospadias is a male birth defect in which the opening of the urethra (the tube through which urine passes) is not located at the tip of the penis as it should be. The milder forms of hypospadias, which are by far the largest number of cases, are relatively easy to repair which usually results in a penis with normal or near-normal function and appearance, and no future problems (135).

A 2004 case-control study from the Netherlands (133) found no link between maternal soy intake and an increased risk of hypospadias. Women who had a soy protein intake of 20 g or more had the exact same risk as those eating no soy (1.0, 0.5-2.2).

A 2000 UK prospective study, of 7,928 boys born to mothers taking part in the Avon Longitudinal Study of Pregnancy and Childhood, found 51 cases of hypospadias (134). There was a trend of more soy meats and soy milk being associated with an increased risk for hypospadias, but the findings were not statistically significant, nor were the results adjusted for any confounding variables.

Two of three studies have found an increased risk of hypospadias if the mothers were vegetarian; more information can be found in Hypospadias and Vegetarian Diets.

Leave a Reply

Your email address will not be published. Required fields are marked *