Does soy have harmful effects on our health?

Last Updated : 31 March 2025

Soy is an important plant-based protein source, particularly for vegetarians, vegans, and those looking to reduce their meat intake. France’s health agency, Agency for Food, Environment, and Occupational Health & Safety (Anses) has recommended to progressively phase out soy and soy-based foods in collective catering menus, such as those in schools, hospitals, and nursing homes.¹ This decision is based on concerns about current intake levels of isoflavones, a type of endocrine active compound abundant in soy, and their potential effects on hormone-related functions. Here are a few things to keep in mind when reading the headlines.

The study behind the headlines

Anses conducted a scientific review to assess the safety of isoflavones, analysing numerous toxicological and epidemiological studies on isoflavone consumption and its human health effects.

Isoflavones are natural compounds found in plants, with soy and soy-based foods like tofu, tempeh, and soy drinks being particularly rich sources. Isoflavones belong to a group of substances called phytoestrogens, meaning they have a structure similar to the hormone oestrogen, which is naturally produced in the human body. Oestrogen is a sex hormone that plays a key role in many bodily functions, including puberty, menstrual cycle, pregnancy, as well as maintaining bone strength.

Because isoflavones are chemically similar to our body’s own oestrogen, they can bind to our oestrogen receptors – the parts of our cells that respond to this hormone. However, their effect isn’t always the same as oestrogen. Depending on the tissue they interact with, their blood dosage or the current blood dosage of other sexual hormones, isoflavones can sometimes mimic oestrogen’s effects and other times block them. In other words, they act like oestrogen in some parts of the body but work against it in others. This versatility feature means that isoflavones can have different effects (beneficial, detrimental, neutral, etc.) depending on a person’s soy intake, age, sex, and overall hormone levels.2,3

Anses reviewed the risks of isoflavone dietary intakes through multiple, standardised steps: First, they defined toxicological reference values (TRVs) – how much isoflavones you can eat daily without it causing harm. This dose was 0.02 mg per kg of body weight for the general population and half as much for vulnerable populations (children before puberty, women of reproductive age, and pregnant women).

To put this into perspective, for a 60 kg adult, the TRV would be 1.2 mg of isoflavones per day (60 kg × 0.02 mg). As an example, as soy drink contains about 7.85 mg of isoflavones per 100 g.4 This means a 60 kg adult could easily exceed their daily limit after drinking around 150 ml of soy drink (about 12 mg of isoflavones in total) – less than a standard glass. However, it’s worth noting that the isoflavone content in soy foods varies significantly among products (even within the same product category), making it harder to predict intake.

Next, the TRVs were compared with dietary records of the French population to see if people’s isoflavone intake might exceed these safety limits. They found that there is a substantial risk of exceeding the TRV among people who ate soy-based foods: 76% of children aged 3 to 5, 53% of girls aged 11 to 17, 47% of men aged 18 and older and women aged 18 to 50 eating soy-based foods exceeded the TRV.

Anses concluded that regular, realistic exposure to soy isoflavones could pose public health risks in France. Therefore, they recommend that: 

• The French population (even more so among vulnerable groups such as children before puberty, women of reproductive age, and pregnant women) limits their daily intake of soy isoflavones
• Collective catering avoids serving isoflavone-rich foods (for all ages).
• Soy food manufacturers and producers take measures to control the isoflavone levels of soybeans and soy products.

This Anses Opinion report does not mean soy is banned in France. It is a precautionary measure to avoid overconsumption and ensure long-term health safety at the individual level.

What to keep in mind when reading the study’s conclusion?

• The toxicological reference values (TRVs) are based on animal studies that are extrapolated to determine safe intake levels for humans.

Anses primarily used data from animal studies, particularly on rats, to define the TRVs for humans. However, rats metabolise isoflavones differently than humans, making it uncertain whether, and at which dosage the same effects would occur in people. There were no human studies robust enough with a dose-response relationship to establish a TRV. To translate findings from animal studies to humans, Anses used a precautionary safety margin, meaning that the established TRV is protective (low) enough to cover for potential differences (absorption, metabolism, excretion, etc.) between animal models and humans.

Learn more about TRVs here : Toxicity reference values (TRVs) | Anses - Agence nationale de sécurité sanitaire de l’alimentation, de l’environnement et du travail

• Anses adopted a precautionary approach due to scientific uncertainties.

The agency applies a conservative stance when assessing the potential risks of isoflavones, particularly because they are endocrine-active substances. Endocrine-active substances like isoflavones can have different effects depending on dose, timing, and life stage. Scientific modelling of toxicological reference values includes significant safety margins to account for unknowns, such as variations in individual sensitivity and long-term effects. This means that the established limits are designed to be highly protective, even in the absence of clear evidence of harm at lower doses on animal models.

• The analysis focuses only on risk assessment, not on health benefits.

The Anses report highlights concerns about isoflavones and hormone-related effects, it does not assess the benefits of soy, soy products or isoflavones themselves. In Europe, risk and benefit assessments are handled separately. Therefore, Anses’ conclusions do not account for any potential positive effects of eating soy.

As a matter of fact, there are currently no officially recognised EU health claim for soy’s benefits. Some studies may suggest that soy intake can reduce the risk of heart disease, improve cholesterol levels, and provide protective effects against certain cancers.^2,5 However, more large-scale epidemiological studies are needed to confirm the benefits. This context helps explain why Anses prioritised caution over potential benefits in its recommendations.

• Isoflavone content varies widely between and within products, but the recommendation applies to all soy foods.

Out of precaution as well as pragmatism, Anses’ recommendation towards catering services does not differentiate between soy products with high and low isoflavone levels, despite significant variability. For example, some soy-based snack biscuits contain 100 times more isoflavones than soy sauce, and even within the same product category, variations of up to twofold have been found in soy desserts on the French market.¹ These variations depend on soybean variety, growing conditions, and processing methods.

What do other authorities say?

• To date, no other official national dietary guidelines (worldwide) advise against overconsumption of soy in general population. To help consumers switch to more plant-based proteins, some countries include soy as part of healthy dietary patterns in national dietary guidelines, such as Austria, Australia, Canada, the Netherlands, the United Kingdom, and the United States.
• The European Food Safety Authority (EFSA) assessed the possible association between the intake of isoflavones from food supplements (daily dose of isoflavones in the range of 35 to 150 mg/day) and harmful effects on the breast, uterus, and thyroid in peri- and post-menopausal women. Their review found no clear evidence that isoflavones increase the risk of breast cancer, affect mammographic density, or alter key biological markers. However, EFSA did not assess dietary native isoflavone intake (from actual soy and soy foods) in the general population.⁶