Isoflavones are a family of plant compounds made mainly by legumes, the bean and pea family. Soybeans hold three main ones: genistein, daidzein and glycitein. Their shape resembles the human hormone estradiol, so they are called phytoestrogens, or plant estrogens.
In the plant they fight microbes and send chemical signals to the nitrogen-fixing bacteria that live in its roots. In soybeans, daidzein and genistein are the signals that switch on the bacteria’s infection genes.
Scientists first noticed the hormone-like effects of isoflavones in the 1940s. Sheep grazing on a type of isoflavone-rich clover had fertility problems. A 1990 review reported that more than 1 million ewes in Australia had permanently damaged reproductive tracts from estrogenic pasture. That does not show what happens in people, and the human studies below are the evidence that counts.
Here are two things most people do not know. First, only about one-third to one-half of people have gut bacteria that turn daidzein into equol, a more estrogen-like compound. Researchers have asked whether those people benefit more from soy.
Second, soy foods and isoflavone supplements are not the same thing. When one lab tested 19 isoflavone supplements, only 5 held at least the amount on the label.
Health benefits of isoflavones
The best-supported findings are about what isoflavones do not do. Meta-analyses (studies that pool the results of many trials) looked at randomized trials, where chance decides who gets the treatment. In them, isoflavones did not change estradiol, FSH (a pituitary hormone) or the uterine lining in postmenopausal women. They did not change testosterone or estrogen levels in men either.
The evidence for benefits is moderate at best. Some meta-analyses found modest drops in hot flashes. A Cochrane review (Cochrane is a group known for careful reviews of trials) found no conclusive effect. Bone results are mixed, and a small drop in LDL cholesterol, the “bad” kind, comes mostly from whole soy foods, not isoflavone pills.
Breast cancer studies point to lower risk in Asian cohorts (large groups followed over time) and no link in Western ones. Many trials used extracts at 54 to 154 mg a day, and each card says whether it tested food or a supplement. Here is what 29 studies found, strongest evidence first, including the trials that found nothing or raised a concern.
Soy and isoflavones did not change men’s testosterone or estrogen
This meta-analysis of 41 clinical studies in adult men covered soy foods, soy protein and isoflavone extracts from soy or red clover. Total testosterone was measured in 1,753 men and free testosterone in 752. Estradiol was measured in 1,000, estrone in 239 and SHBG, a protein that carries sex hormones, in 967. No outcome changed, whatever the isoflavone dose or study length.
Soy isoflavones did not change hormones or the uterine lining after menopause
In 40 randomized trials, 3,285 postmenopausal women took a median 75 mg of isoflavones a day for a median 24 weeks, or a non-isoflavone control. Estradiol, FSH and vaginal maturation did not change, and lining thickness moved only -0.22 mm, a gap small enough to be chance. The authors rate the certainty of the evidence high to moderate. One author works for the Soy Nutrition Institute Global, which is funded by the United Soybean Board and soy industry members.
Isoflavone extracts cut hot flash frequency 20.6% and severity 26.2% against placebo
This meta-analysis pooled randomized trials of soy isoflavones, extracted or man-made, against a placebo (dummy pill), all double-blind: nobody knew who got what. The median dose was 54 mg a day, for 6 weeks to 12 months; 13 trials measured frequency and 9 measured severity. Results varied a lot between trials. Supplements with more than 18.8 mg of genistein, the median across studies, were more than twice as potent at cutting frequency.
Cochrane review: no conclusive proof phytoestrogens ease hot flashes or night sweats
This Cochrane review pooled 43 randomized trials in 4,364 women who used foods or supplements rich in phytoestrogens for at least 12 weeks. Five pooled trials of Promensil, a red clover extract, showed -0.93 flashes a day against placebo, a gap that could be chance. Four trials of extracts with more than 30 mg a day of genistein consistently reduced flashes, but placebo responses were strong (1% to 59%). There was no sign of stimulation of the uterine lining or vagina for up to 2 years.
Show 25 more studies
Phytoestrogens modestly cut hot flashes by 1.31 a day, not night sweats
This meta-analysis pooled 62 studies in 6,653 women. Soy isoflavones alone, from food or supplements, cut hot flashes by 0.79 a day and vaginal dryness scores by 0.26. Night sweats did not clearly change. 46 of the 62 trials (74%) had a high risk of bias, meaning flaws that can skew results, in 3 or more areas.
Equol, made by some gut bacteria, lowered hot flash scores
This meta-analysis pooled 5 randomized trials in 728 women. They tested equol itself, or soy isoflavones in women whose gut bacteria make equol from daidzein. Three of the five trials found a clear benefit and two did not, and bias risk was generally low. See “How isoflavones work” for why only some people make equol.
Red clover cut hot flashes by 1.73 a day, but results varied
This meta-analysis pooled 8 randomized trials (10 comparisons) of red clover against placebo; red clover is a different isoflavone source from soy. The likely effect ranged from -3.28 to -0.18 flashes a day. Benefit was clearer in women with 5 or more flashes a day, at doses of 80 mg a day or more, and after 12 weeks. It was also clearer with products higher in biochanin A, and the authors say more well-designed trials are needed.
Soy isoflavones raised spine bone density 20.6 mg/cm² in menopausal women
This meta-analysis pooled 10 randomized trials in 608 women. The gain was larger above 90 mg a day (28.5 mg/cm²) and with 6 months of treatment (27 mg/cm²). Spine bone mineral content rose 0.93 g, a borderline result that could be chance. Fractures were not measured.
In Western women, soy isoflavones did not clearly change spine bone density
This systematic review covered 17 trials, and 12 of them, with 1,433 women, measured lower-spine density. The difference, 9.86 mg/cm², was small enough to be chance. Leaving out the one genistein study, isoflavone mixtures gave 0.73, and more dose or longer treatment did not help. The authors say isolated genistein, and how well each person processes isoflavones, remain open questions.
Soy isoflavones for 3 years did not slow bone loss after menopause
In this double-blind trial, 224 postmenopausal women took 80 or 120 mg of soy isoflavones a day, as Novasoy tablets donated by Archer Daniels Midland. Everyone also got 500 mg of calcium and 600 IU of vitamin D3. Spine density fell about 2% (-2.08%) whatever the treatment, and only 120 mg showed a modest benefit at the hip (femoral neck), after adjustment. Adverse events, uterine lining thickness and bone markers did not differ.
120 mg a day slowed whole-body bone loss, not hip or spine
In this double-blind trial, 403 postmenopausal women took 80 or 120 mg a day of soy hypocotyl isoflavones, or a placebo, for 2 years. All also took calcium and vitamin D. On 120 mg, whole-body density fell less than on placebo, at both 1 year and 2 years. Density at single sites, such as the hip and spine, and bone markers did not differ.
Genistein for 2 years raised spine and hip bone density
This double-blind trial in Italy gave 389 women with osteopenia (low bone density) 54 mg of genistein a day or a placebo. All also took calcium and vitamin D. It was isolated genistein, not a soy food or a mixture, and spine density rose 0.049 g/cm² on it and fell 0.053 on placebo. The uterine lining did not change in thickness, but stomach and bowel side effects were more common (19% vs 8%), and fractures were not measured.
A year of isoflavone soy protein did not help bone or cholesterol
This double-blind Dutch trial had 202 healthy postmenopausal women aged 60 to 75; 175 were analyzed. Daily, they took milk protein or 25.6 g soy protein with 99 mg of isoflavones: 52 mg genistein, 41 mg daidzein and 6 mg glycitein. Blood genistein confirmed they took them (615.1 vs 17.2 nmol/L), but bone density, cholesterol and thinking did not improve. The authors say this applies to women starting at age 60 or later.
Soy isoflavones lowered LDL by 0.13 mmol/L (about 5 mg/dL)
This meta-analysis pooled 11 randomized trials. LDL fell 3.58% and total cholesterol 0.10 mmol/L (millimoles per liter, a unit for blood levels), while HDL and triglycerides did not change. Soy protein with the isoflavones removed also lowered LDL (0.10 mmol/L), so the protein and the isoflavones cannot be separated. Drops were larger in people with high cholesterol.
Soy protein lowered LDL by 4.76 mg/dL across 41 trials
This meta-analysis drew on the 46 trials the FDA, the US food regulator, found when it weighed revoking the soy protein heart health claim. 43 gave data, with a median 25 g of soy protein a day for 6 weeks against non-soy protein. LDL fell about 3% to 4%, and total cholesterol fell 6.41 mg/dL. About 75% of trials showed a drop but only a minority were clear on their own, and this tested soy protein, not isolated isoflavones.
Whole soy foods cut LDL most; isoflavone supplements had no effect
This meta-analysis pooled 35 randomized studies (50 comparisons) of soy products lasting 4 weeks to 1 year. Whole soy foods (soy milk, soybeans and nuts) lowered LDL by 11.06 mg/dL, and soy extracts by 3.17 mg/dL. Isoflavone supplements had no effect, and overall LDL fell 4.83 mg/dL and triglycerides 4.92, while HDL rose 1.40. Results varied a great deal between trials.
3 years of soy protein did not thicken the uterine lining
In this double-blind trial, 350 postmenopausal women were randomized to milk protein or soy protein with 154 mg of isoflavones a day. 224 women with a uterus were analyzed, and lining thickness and rates of hyperplasia (overgrowth of the lining) did not differ. All 9 biopsies on soy were benign, while one woman on placebo had complex hyperplasia with atypia; her hysterectomy showed stage IB endometrial cancer. The trial is small for rare events.
Isoflavones did not change breast density after menopause, slightly raised it before
This meta-analysis pooled 8 randomized trials in 1,287 women, lasting 6 months to 3 years. Density changed +0.69% in all women and -1.10% after menopause, both small enough to be chance. Before menopause it rose 1.83%, a result lost in one of three sensitivity checks. The authors call for longer trials to see if these small shifts matter.
5 years of soy isoflavones: uterine overgrowth in 3.37%, placebo 0%
In this double-blind trial, 376 postmenopausal women with a uterus took 150 mg of soy isoflavones a day or a placebo, and 298 finished. Biopsies found no cancers; hyperplasia is a thickening of the uterine lining. 70% of women on soy had a lining classed as atrophic or not assessable, against 81% on placebo. In 2025 the journal added an expression of concern, with no reason given in PubMed, and the meta-analysis above found no change in lining thickness.
16 mg of phytoestrogens went with 6 women developing overt hypothyroidism
This double-blind crossover trial (everyone tried both doses, in turn) had 60 patients with subclinical hypothyroidism, a mildly underactive thyroid. Each took 30 g of soy protein with 2 mg or 16 mg of phytoestrogens for 8 weeks, with an 8-week break between. All 6 who progressed did so after the 16 mg phase (3.6 times the rate, range 1.9 to 6.2). On the same dose blood pressure, insulin resistance and CRP (a marker of inflammation) fell, while blood fats did not change.
66 mg of soy phytoestrogens did not change thyroid hormones
This double-blind crossover trial had 44 patients with subclinical hypothyroidism. They took 30 g of soy protein with 66 mg of phytoestrogens, and 30 g with none, for 8 weeks each. There was an 8-week break between, and 2 patients progressed to overt hypothyroidism after the high dose. TSH, free thyroxine and T3 did not differ, and the same research group ran the 2011 trial above.
57 days of soy isoflavones did not change semen quality
Healthy young men each took milk protein, low-isoflavone soy protein isolate (1.64 mg a day) and high-isoflavone isolate (61.7 mg a day). In this randomized crossover trial, each lasted 57 days, with 28-day breaks between. Volume, sperm concentration, count, movement and shape did not change.
High soy: lower breast cancer odds in Asian women, not Western
This meta-analysis pooled 35 observational studies, which watch what people eat rather than assign it. In Asian countries, high isoflavone intake went with 0.59 times the odds of breast cancer, before and after menopause, though this differed by study design. In Western postmenopausal women the figure was 0.92, with no clear link once split by design. Observational data cannot show that soy caused the difference.
Moderate soy was not linked to breast cancer in 300,000 Chinese women
The China Kadoorie Biobank followed women aged 30 to 79, who ate an average 9.4 mg of isoflavones a day. Over 10 years, 2,289 developed breast cancer. The top quarter of intake (19.1 mg a day) had the same risk as the bottom quarter (4.5 mg a day): 1.00. A pooled analysis of studies that follow people forward found 3% lower risk for each 10 mg a day (range 1% to 5%).
In breast cancer survivors, more soy went with fewer deaths and recurrences
The Shanghai Breast Cancer Survival Study followed 5,042 Chinese survivors for a median 3.9 years, with 444 deaths and 534 recurrences or breast cancer deaths. The highest soy protein intake went with 0.71 times the risk of death and 0.68 times the risk of recurrence. It held for both estrogen receptor types and with or without tamoxifen; four-year recurrence was 11.2% and 8.0% in the lowest and highest quarters. One author received a small research fund from the United Soybean Board, which had no role in the study.
Soy before diagnosis went with better survival; soy after, no clear link
This meta-analysis of 12 studies covered 37,275 women with breast cancer. Soy before diagnosis went with 0.84 times the risk of death. After diagnosis, overall survival (0.80) and breast cancer survival (0.83) showed no clear link, but both pre- and post-diagnosis intake went with less recurrence. The authors call the evidence limited.
In 99 clinic men, high soy went with fewer sperm
This cross-sectional study, a single snapshot, asked male partners of subfertile couples at Massachusetts General Hospital about 15 soy foods over the previous 3 months. Against men who ate no soy, the gap was 41 million sperm/mL (range -74 to -8), and larger in overweight or obese men. Motility, shape and volume were unrelated. It is a small study of one clinic, and the randomized trial above found no change in healthy men.
In animals, isoflavones shrank tumor volume and weight but grew tumor area
This meta-analysis pooled 30 animal studies of breast, ovarian and uterine cancer models, not people. Tumor volume (Hedges g -1.151) and tumor weight (-2.584) fell, while tumor area rose (1.136). The authors advise caution for patients with hormone-dependent cancers, particularly with supplements.
In mice, low-dose genistein cancelled tamoxifen’s effect on breast tumors
This study used mice carrying human estrogen-receptor-positive breast cancer cells (MCF-7), not people. Genistein at 250 and 500 ppm in the diet cancelled tamoxifen, while 1000 ppm did not. The authors advise caution for patients taking tamoxifen. The Shanghai survivor cohort above did not see worse outcomes in tamoxifen users.
Foods high in isoflavones
Isoflavones come almost entirely from legumes, and soybeans and soy foods are far ahead of everything else. NutriScrolls has no USDA ranking for isoflavones, so the table below shows published measurements. They come from USDA’s isoflavone database (release 2.0, 2008), which adds up daidzein, genistein and glycitein as aglycones (the sugar-free form), and two papers. The ranges are wide because soybean variety, crop year and processing all change the amount.
By that arithmetic, 54 mg, the median dose in the hot flash trials above, is about 35 g of raw soybeans. It is also about 90 g of tempeh, 175 g of firm tofu or 500 g of plain soy milk. Chickpeas, the legume most people know, hold under 1 mg of the three main soy isoflavones per 100 g. Red clover, the other main source in supplements, is not a common food.
| Food | Amount per 100 g | Source |
|---|---|---|
| Soybeans, mature seeds, raw | 154.53 mg (range 10.04 to 440.72) | 1 |
| Soy flour, defatted | 150.94 mg | 1 |
| Natto (fermented soybeans) | 82.29 mg (46.40 to 124.10) | 1 |
| Tempeh | 60.61 mg (6.88 to 179.20) | 1 |
| Edamame (green soybeans, raw) | 48.95 mg (0.05 to 120.94) | 1 |
| Miso | 41.45 mg (3.26 to 99.52) | 1 |
| Tofu, firm (made with calcium sulfate and magnesium chloride) | 30.41 mg (3.12 to 142.30) | 1 |
| Tofu, silken | 18.04 mg (8.10 to 26.90) | 1 |
| Soy milk, original and vanilla | 10.73 mg (1.10 to 31.03) | 1 |
| 47 soy foods sold in Hong Kong | 1 to 80 mg. Mean 9.99 mg in the soy milk group and 40 mg in the bean curd skin group | 2 |
| Chickpeas, raw | 0.38 mg (0.00 to 1.06) of daidzein, genistein and glycitein. Also 1.54 mg of biochanin A and 0.12 mg of formononetin | 1 |
| Germinated chickpea seeds | Isoflavone content rose more than 100-fold with germination, to about 5 times the total in germinated soybeans (no mg figure in the abstract) | 3 |
Sources
- USDA Database for the Isoflavone Content of Selected Foods, Release 2.0. Nutrient Data Laboratory, Agricultural Research Service, US Department of Agriculture. 2008. Mean mg per 100 g of edible portion, with the lowest and highest sample in brackets. Defatted soy flour is the mean given in the documentation.
- Chan SG, Murphy PA, Ho SC, et al. Isoflavonoid content of Hong Kong soy foods. J Agric Food Chem. 2009. In aglycone equivalents per 100 g of wet weight.
- Wu Z, Song L, Feng S, et al. Germination dramatically increases isoflavonoid content and diversity in chickpea (Cicer arietinum L.) seeds. J Agric Food Chem. 2012.
How isoflavones work
Most of what we know about how isoflavones act comes from test tubes, cells and animals. Three ideas come up again and again. Isoflavones fit the body’s estrogen receptors, with a preference for one of the two types, and they can block some enzymes in the lab. Gut bacteria also change them into other compounds, which is one reason people respond differently.
Here is each one, and what has been seen in people.
They fit estrogen receptors, but not like estradiol
Estrogen works through two main receptors, called alpha and beta. In a 1998 test-tube study, genistein competed with estradiol more strongly at the beta receptor than at the alpha receptor. It switched on both receptors at 1 to 10 nM (nanomolar, a tiny concentration). Industrial estrogen-like chemicals such as bisphenol A needed 100 to 1000 nM, and the study found no blocking (antiestrogen) activity for the isoflavones it tested.
A schematic of the ranking in the study. The abstract gives the order of binding, not one number for each receptor, so the bar lengths are not to scale.
Whether this means isoflavones act like estrogen in people is a different question. The 2025 meta-analysis of 40 trials in the study cards above found no change in estradiol, FSH or the uterine lining in postmenopausal women. Its authors say isoflavones likely act as selective estrogen receptor modulators, which differ clinically from the hormone itself. One author is employed by an organization funded by the soy industry.
Gut bacteria change them
Bacteria in the large intestine rework isoflavones. Daidzein can become equol, which has a selective pull toward estrogen receptor beta. Only some adults make it: a review put the share at about one-third to one-half of people. Gut bacteria also reduce two other plant isoflavones, formononetin and biochanin A, to daidzein and genistein, says the documentation of the USDA isoflavone database.
The conversions are drawn as a flow, and the bar is drawn to scale. The review says “approximately one-third to one-half” of people make equol.
Whether it matters has been tested in people. A pooled analysis of 5 trials found that equol, or soy isoflavones in equol producers, lowered hot flash scores. One trial gave 160 Japanese women who do not make equol a 10 mg a day S-equol supplement for 12 weeks. It cut hot flashes by 1.9 a day, against 1.0 a day on placebo.
They block some enzymes in the lab
In 1987, genistein was shown to block tyrosine kinases, enzymes that pass growth signals inside cells, in test tubes and cultured cells. In rats, genistein in the feed cut the activity of thyroid peroxidase, an enzyme needed to make thyroid hormone, by up to 80%. Rats on standard soy-based chow had about 50% lower activity than rats on soy-free chow. Even so, their thyroid hormone levels, thyroid weights and tissue did not differ.
Drawn to scale, with soy-free feed as 100. “About 50” and “as low as 20” come from the study’s statements that activity was about 50% lower on soy chow and reduced by up to 80% on genistein.
The authors suggest the remaining enzyme was enough to keep the thyroid steady when nothing else disturbed it. Whether that holds in people with a thyroid problem is what the two trials in subclinical hypothyroidism above tested, with different results.
What has been seen in people
The lab effects are small in people. Isoflavone extracts cut hot flash frequency by 20.6% against placebo in a meta-analysis, but a Cochrane review found no conclusive evidence. Soy protein lowered LDL cholesterol by 4.76 mg/dL, and whole soy foods did better than isoflavone pills. Bone results were mixed, hormone levels did not change, and all of these are among the study cards above.
Sources
- Kuiper GG, Lemmen JG, Carlsson B, et al. Interaction of estrogenic chemicals and phytoestrogens with estrogen receptor beta. Endocrinology. 1998.
- Viscardi G, Back S, Ahmed A, et al. Effect of soy isoflavones on measures of estrogenicity: a systematic review and meta-analysis of randomized controlled trials. Adv Nutr. 2025. One author is employed by the Soy Nutrition Institute Global.
- Yuan JP, Wang JH, Liu X. Metabolism of dietary soy isoflavones to equol by human intestinal microflora: implications for health. Mol Nutr Food Res. 2007.
- Setchell KD, Clerici C. Equol: history, chemistry, and formation. J Nutr. 2010.
- USDA Database for the Isoflavone Content of Selected Foods, Release 2.0. Nutrient Data Laboratory, Agricultural Research Service, US Department of Agriculture. 2008. Documentation, section “Other phytoestrogens”.
- Aso T, Uchiyama S, Matsumura Y, et al. A natural S-equol supplement alleviates hot flushes and other menopausal symptoms in equol nonproducing postmenopausal Japanese women. J Womens Health (Larchmt). 2012.
- Daily JW, Ko BS, Ryuk J, et al. Equol decreases hot flashes in postmenopausal women: a systematic review and meta-analysis of randomized clinical trials. J Med Food. 2019.
- Akiyama T, Ishida J, Nakagawa S, et al. Genistein, a specific inhibitor of tyrosine-specific protein kinases. J Biol Chem. 1987.
- Chang HC, Doerge DR. Dietary genistein inactivates rat thyroid peroxidase in vivo without an apparent hypothyroid effect. Toxicol Appl Pharmacol. 2000.
Is isoflavones safe? Supplements, dose and interactions
In the controlled trials above, soy or isoflavone intake did not change hormone levels in men or in postmenopausal women. For supplements, EFSA, the European food safety agency, reviewed isoflavones taken by postmenopausal women in 2015. It said the evidence reviewed “does not suggest there are harmful effects on the three organs considered”: the breast, uterus and thyroid.
EFSA limited that to the doses and intake periods studied, about 3 months to just under a year. It found no data for women in the menopausal transition or with a personal or family history of cancer.
In the US, the FDA authorized a health claim in 1999. It says 25 g of soy protein a day, in a diet low in saturated fat and cholesterol, may reduce heart disease risk. On October 31, 2017 the FDA published a proposed rule to revoke that claim, and later extended the comment period to March 19, 2018. The current Code of Federal Regulations still contains the claim, and we found no final rule.
- Trial doses and side effects. The hot flash trials used a median of 54 mg a day. Bone and uterus trials used 80 to 154 mg a day for up to 5 years. In the 2-year genistein bone trial, stomach and bowel side effects were more common (19% vs 8%). The Cochrane review saw no sign of stimulation of the uterine lining or vagina for up to 2 years. The one 5-year trial that found more hyperplasia (3.37% vs 0%) now carries an expression of concern from its journal.
- Supplement quality varies. Of 19 isoflavone preparations tested in one lab, only 5 held at least the labeled amount. In another test of 11 soy and red clover supplements, soy-based products held less than labeled once measured as aglycone equivalents. Trials named in the cards used Novasoy tablets from Archer Daniels Midland and the red clover extract Promensil. Check what a product measured, not just what it claims.
- Breast cancer and tamoxifen. The Shanghai cohort found no harm from soy foods in tamoxifen users. But a mouse study found low-dose genistein cancelled tamoxifen, and EFSA had no data for women with a cancer history. A meta-analysis of animal studies advises caution with supplements, so ask your oncology team before taking them. This concern is about supplements and mice more than food.
- Thyroid medicine. One trial linked 16 mg a day of soy phytoestrogens to overt hypothyroidism in 6 people with a thyroid problem. A later trial at 66 mg found no change. A systematic review found limited evidence that soy lowers absorption of levothyroxine, a thyroid hormone medicine. In a case report, a patient needed a higher dose while taking a soy supplement. If you take levothyroxine, ask your clinic about timing.
How to get more isoflavones
- Best food. Natto holds 82.29 mg per 100 g, tempeh 60.61 mg and edamame 48.95 mg. Miso holds 41.45 mg and firm tofu 30.41 mg, against 10.73 mg for soy milk.
- Food or extract. In a pooled analysis of 35 studies, whole soy products lowered LDL by 11.06 mg/dL and extracts by 3.17 mg/dL. Isoflavone supplements had no effect on blood fats. The hot flash trials, however, used extracts, at a median of 54 mg a day.
- Processing changes the amount. USDA’s documentation says alcohol extraction in processing lowers isoflavone content significantly. 47 soy foods sold in Hong Kong ranged from 1 to 80 mg per 100 g. Chickpea seeds that had germinated held more than 100 times the isoflavones of ungerminated seeds.
Sources
- Isoflavones in food supplements for post-menopausal women: no evidence of harm. European Food Safety Authority, news story, October 21, 2015. EFSA Panel on Food Additives and Nutrient Sources added to Food (ANS), opinion in EFSA Journal 2015;13(10):4246.
- FDA extends comment period on proposed rule to revoke authorized health claim regarding soy protein and coronary heart disease. US Food and Drug Administration, constituent update, January 12, 2018. Docket FDA-2017-N-0763.
- 21 CFR 101.82: Health claims: soy protein and risk of coronary heart disease (CHD). Code of Federal Regulations, US Food and Drug Administration. Added October 26, 1999 (64 FR 57732).
- Reiter E, Beck V, Medjakovic S, et al. Comparison of hormonal activity of isoflavone-containing supplements used to treat menopausal complaints. Menopause. 2009.
- Andres S, Hansen U, Niemann B, et al. Determination of the isoflavone composition and estrogenic activity of commercial dietary supplements based on soy or red clover. Food Funct. 2015.
- Alekel DL, Van Loan MD, Koehler KJ, et al. The soy isoflavones for reducing bone loss (SIRBL) study: a 3-y randomized controlled trial in postmenopausal women. Am J Clin Nutr. 2010. Tablet supplier named in the full text.
- Du M, Yang X, Hartman JA, et al. Low-dose dietary genistein negates the therapeutic effect of tamoxifen in athymic nude mice. Carcinogenesis. 2012.
- Chakravarti B, Rajput S, Srivastava A, et al. A systematic review and meta-analysis of the effects of dietary isoflavones on female hormone-dependent cancers for benefit-risk evaluation. Phytother Res. 2024.
- Wiesner A, Gajewska D, Paśko P. Levothyroxine interactions with food and dietary supplements: a systematic review. Pharmaceuticals (Basel). 2021.
- Bell DS, Ovalle F. Use of soy protein supplement and resultant need for increased dose of levothyroxine. Endocr Pract. 2001.
- Tokede OA, Onabanjo TA, Yansane A, et al. Soya products and serum lipids: a meta-analysis of randomised controlled trials. Br J Nutr. 2015.
- USDA Database for the Isoflavone Content of Selected Foods, Release 2.0. 2008.
- Chan SG, Murphy PA, Ho SC, et al. Isoflavonoid content of Hong Kong soy foods. J Agric Food Chem. 2009.
- Wu Z, Song L, Feng S, et al. Germination dramatically increases isoflavonoid content and diversity in chickpea seeds. J Agric Food Chem. 2012.
History of isoflavones
Genistein takes its name from the dyer’s broom plant, Genista tinctoria, the plant it was first isolated from in 1899. Its structure was established in 1926, and it was made in the lab in 1928. Daidzein and genistein were later found to be the main isoflavones in soybeans.
The hormone-like effects of isoflavones came to the attention of scientists in the 1940s, when fertility problems appeared in sheep grazing on isoflavone-rich clover. Over the following decades the problem in Australian flocks was traced to plant estrogens in pasture. By 1990, more than 1 million ewes there were reported to have permanently damaged reproductive tracts.
In the 1950s isoflavones were tested as growth promoters in animal feed. Wide interest in soy and human health began in the 1990s, partly through research sponsored by the US National Cancer Institute. In 1995 soy protein drew worldwide attention for lowering cholesterol, and the FDA authorized its heart health claim in 1999.
Equol, the gut product of daidzein, was first isolated from horse urine in 1932 and found in human urine 50 years later. In 2002 it was proposed that people who make equol are more likely to benefit from soy. In 2017 the FDA proposed to revoke the soy protein claim. We found no final decision.
Sources
- Messina M. A brief historical overview of the past two decades of soy and isoflavone research. J Nutr. 2010.
- Adams NR. Permanent infertility in ewes exposed to plant oestrogens. Aust Vet J. 1990.
- Setchell KD, Clerici C. Equol: history, chemistry, and formation. J Nutr. 2010.
- Yang Q, Wang G. Isoflavonoid metabolism in leguminous plants: an update and perspectives. Front Plant Sci. 2024.
- Rolfe BG. Flavones and isoflavones as inducing substances of legume nodulation. Biofactors. 1988.
- 21 CFR 101.82: Health claims: soy protein and risk of coronary heart disease (CHD). Code of Federal Regulations, US Food and Drug Administration.
- Genistein. Wikipedia.
Frequently asked questions about isoflavones
What do isoflavones do for the body?
Isoflavones are plant compounds that can attach to the body’s estrogen receptors. In 40 trials in postmenopausal women, they did not change estradiol, FSH or the uterine lining. Measured benefits are modest: 20.6% fewer hot flashes in one meta-analysis, a 4.76 mg/dL LDL drop from soy protein, and mixed results for bone. No benefit reaches strong evidence.
Do isoflavones help with hot flashes?
A little, in some analyses. One meta-analysis of extracts found frequency fell 20.6% and severity 26.2%. A JAMA analysis of 62 studies found soy isoflavones cut 0.79 hot flashes a day. But a Cochrane review of 43 trials found no conclusive evidence, with placebo responses of 1% to 59%.
Extracts with more than 30 mg a day of genistein looked most promising.
Which foods are highest in isoflavones?
Soybeans and the foods made from them. Per 100 g, USDA release 2.0 gives raw soybeans 154.53 mg, natto 82.29 mg and tempeh 60.61 mg. Edamame has 48.95 mg, miso 41.45 mg, firm tofu 30.41 mg and soy milk 10.73 mg. Chickpeas hold 0.38 mg of the three main soy isoflavones.
Are isoflavones safe if you have had breast cancer?
The evidence is mixed, so ask your cancer team. In 5,042 survivors in Shanghai, the highest soy protein intake went with lower death (0.71 times the risk) and recurrence (0.68), including in tamoxifen users. A meta-analysis found no clear link between soy after diagnosis and survival. EFSA found no data on supplements for women with a cancer history, and a mouse study found genistein cancelled tamoxifen.
Supplements are a bigger unknown than soy foods.
Do soy isoflavones affect the thyroid?
The trials we found tested people with an underactive thyroid, and they disagree. One found 6 women progressed to overt hypothyroidism after 16 mg of soy phytoestrogens (3.6 times the rate). A later trial at 66 mg found no change in thyroid hormones. EFSA found no thyroid harm in postmenopausal women at supplement doses, and soy may lower the absorption of levothyroxine, so ask your clinic about timing.
Do isoflavones lower testosterone in men?
No. A meta-analysis of 41 studies found no effect of soy foods, soy protein or isoflavone extracts on testosterone, estradiol or estrone. Dose and length made no difference. A crossover trial of 61.7 mg a day for 57 days found no change in semen quality. A clinic study of 99 men linked the highest soy intake to 41 million fewer sperm per milliliter, but that cannot show cause.
How much isoflavone should you take a day?
No dose is proven best. The hot flash trials had a median of 54 mg a day, and bone and uterus trials used 80 to 154 mg. EFSA’s 2015 review covered supplements taken for about 3 months to just under a year. About 100 g of firm tofu supplies roughly 30 mg, so food alone can reach the lower end.
What is equol, and do you need to make it?
Equol is a compound that some gut bacteria make from daidzein. About one-third to one-half of people make it. A pooled analysis of 5 trials found equol, or soy isoflavones in equol producers, lowered hot flash scores. Whether being a producer changes the benefit of soy foods is still being studied.