Are lectins bad for you?
The claim: Lectins in beans, grains and tomatoes damage your gut.
Last checked 9 October 2026
Raw beans really can poison you. Cooked properly, most of their lectins are gone, and we found no good evidence that cooked lectins harm healthy people.
Lectins are proteins plants use as a defense. In raw or undercooked beans they can cause severe vomiting and diarrhea. Europe's food safety authority, EFSA, found no health concern from bean lectin fully deactivated by soaking and boiling, but a health concern in its scenario of undercooked beans that keep part of it active. One lab study found no lectin activity left in soaked and boiled kidney beans and black beans, or in boiled lentils; chickpeas showed none in that test, but a second version of the test still found some after an hour's boiling. Low-heat slow cooking is not enough, and can make raw beans' lectins more active.
What lectins are, and where they are
Harvard's Nutrition Source explains why lectins get attention:
Lectins are defined as proteins that bind to carbohydrates. The same features that lectins use to defend plants in nature may cause problems during human digestion.
While pulses (e.g., green beans, red kidney beans, jack beans) and wheat germ contain comparatively high amounts, nightshade plants (e.g., potatoes, tomatoes) and certain fruits (e.g., bananas) contain only small amounts of lectins.
Pulses are dried beans, lentils and peas. The second quote is from Germany's Federal Institute for Risk Assessment (BfR), which reviewed lectins in 2025.
The real danger: raw and undercooked beans
Europe's food safety authority, EFSA, published its assessment of lectins in food in January 2026. It could assess only one lectin, phytohaemagglutinin (PHA), found in Phaseolus beans such as kidney, black and lima beans. EFSA uses a margin of exposure (MOE), which compares a dose that affected lab animals with what people might eat; the smaller the margin, the bigger the concern:
Based on the available evidence, the CONTAM Panel considered only phytohaemagglutinin (PHA), a legume lectin from beans (Phaseolus sp.), for the risk characterisation.
Consumption of partially processed red and white kidney beans can cause GI symptoms, namely, nausea, vomiting, abdominal pain and diarrhoea.
An arbitrary acute exposure scenario, where only 50% of the lectins are inactivated due to insufficient cooking of food containing lectins (e.g. beans), would result in MOEs below 100. The Panel, accounting for the uncertainties affecting the exposure and hazard assessments, concluded with at least 95% probability that such a dietary exposure would raise health concerns.
However, ingesting certain lectins, which are mainly found in raw pulses, can lead to health impairments, including symptoms such as nausea, vomiting and diarrhoea. Children are particularly at risk due to their low body weight.
Between July 1976 and February 1989, 50 incidents of suspected red kidney bean poisoning were reported in the UK.
In many of the outbreaks reported the implicated beans were consumed raw or following an inadequate heat process.
The BfR also describes a Japanese television programme that promoted powdered, roasted white kidney beans:
One thousand viewers suffered from acute intestinal problems and 100 had to be hospitalised.
The BfR says the roasting had not heated the beans enough to destroy their lectins.
What cooking does, measured
A Danish study measured lectin activity in foods raw and after the soaking and boiling shown. Higher numbers mean more active lectin:
| Food | Raw | After preparation | Preparation | Test |
|---|---|---|---|---|
| Black beans | 26,429 ± 5,603 | not detected | soaked 12 h, boiled 30 min | NTT |
| Lima beans | 26,526 ± 5,624 | not detected | soaked 12 h, boiled 30 min | NTT |
| Red kidney beans | 13,214 ± 2,828 | not detected | soaked 12 h, boiled 30 min | NTT |
| Green beans | 6,656 ± 1,411 | not detected | boiled 10 min | NTT |
| Lentils (red) | 3,328 ± 706 | not detected | boiled 10 min | NTT |
| Lentils (brown) | 3,328 ± 706 | not detected | boiled 20 min | NTT |
| Soybeans (dry) | 3,328 ± 706 | not detected | soaked 12 h, boiled 45 min | NTT |
| Edamame | 416 ± 88 | not detected | boiled 5 min | NTT |
| Chickpeas (brown) | not detected | not detected | soaked 12 h, boiled 1 h | NTT |
| Chickpeas (yellow) | not detected | not detected | soaked 12 h, boiled 1 h | NTT |
| Chickpeas (brown) | 13,312 ± 2,822 | 6,656 ± 1,411 | soaked 12 h, boiled 1 h | TT |
| Chickpeas (yellow) | 13,312 ± 2,822 | 13,312 ± 2,822 | soaked 12 h, boiled 1 h | TT |
| Potato | 826 ± 175 | not detected | boiled 20 min | NTT |
| Tomato | 256 ± 54 | 104 ± 22 | boiled 5 min | NTT |
| Red bell pepper | 26 ± 6 | not detected | boiled 5 min | NTT |
| Eggplant | not detected | not detected | boiled 10 min | NTT |
| Rice (type not stated) | 208 ± 44 | not detected | boiled 15 min | TT |
| Wheat (grain) | not detected | not detected | boiled 15 min | NTT |
"Not detected" means the test found no activity, not that every trace is gone. Rows marked TT come from a second version of the test (trypsin-treated), so they can't be compared with the NTT rows. Chickpeas were tested both ways: none in the NTT test, raw or boiled, but activity in the TT test. The study lists its wheat sample only as wheat; the BfR's "comparatively high" figure is for wheat germ. The study's own summary:
The results showed the highest lectin activity in species from the Fabaceae family and demonstrated that soaking and boiling have an effect on the levels of active lectins.
Fabaceae is the legume family: beans, lentils, peas, soy and peanuts.
Slow cookers and low heat
Heat destroys bean lectins, but only if it is hot enough. The BfR describes a test on red beans:
the lectins in some samples were largely destroyed after just 2 minutes at 100 °C, while it took almost eight times as long at 90 °C. At a heating temperature of 80 °C, there was a considerable increase in the agglutination force (clumping force), which reached a maximum after 15 minutes that was seven times higher than the initial value.
However, raw beans simmered at low heat such as in a slow-cooker or undercooking the beans will not remove all the lectins.
Since the lectin activity is destroyed by high heat, the following applies: cook fresh beans in boiling water for at least 30 minutes. Dried pulses should be soaked for at least five hours and then boiled in fresh water, after pouring away the soaking water.
Soaking (6–12 h) combined with heat treatment (≥ 100°C, > 30 min) are highly effective conditions for deactivating lectins from plant sources (e.g. PHA present in kidney beans). However, since the effect of the soaking time can vary depending on the kidney bean varieties, it is advisable to continue soaking until the beans have softened before cooking.
Harvard and the BfR differ on stewing: Harvard counts it among the wet high-heat methods that work, while the BfR calls steaming and stewing unsuitable for most pulses. Both agree that raw beans in a low-heat slow cooker are not safe.
Do cooked lectins harm your gut?
This is the claim behind lectin-free diets. Part of the mechanism is real:
Lectins can bind to carbohydrate structures of cells in the human intestine or blood and thereby induce cell division, cause agglutination ("clumping") of red blood cells (so-called haemagglutination), affect the immune system, alter the permeability of the intestinal barrier and generally disrupt cell metabolism
Because lectin proteins bind to cells for long periods of time, they can potentially cause an autoimmune response and are theorized to play a role in inflammatory conditions like rheumatoid arthritis and type 1 diabetes.
However, there is very limited research in humans on the amount of active lectins consumed in the diet and their long-term health effects.
The available epidemiological data on GI and metabolic effects are limited both in terms of volume and validity and cannot be used further for risk characterisation.
Those effects are for active lectins. In the lab test above, soaked and boiled beans and boiled lentils showed none, though chickpeas kept some in the second version of the test. Some lectins do survive eating: EFSA reports active peanut and wheat germ lectins reaching the blood (see the peanut card below). EFSA's view on lectins fully deactivated by proper soaking and boiling, and the limit of what it could assess:
The Panel also noted that exposure to completely deactivated lectins in food prepared following adequate food processing practices (e.g. soaking and boiling) would not raise health concerns. No risk characterisation could be performed for other lectins due to the lack of relevant toxicological data and/or in some cases lack of occurrence data.
Harvard also points to large population studies, in which lectin-containing foods like legumes, whole grains and nuts go with lower rates of cardiovascular disease and type 2 diabetes (see the last card below). Associated is not the same as caused: people who eat more beans and whole grains may differ in other ways too.
What you'll hear at the table
The same arguments come up whenever lectins are argued over. Here is each one, checked against its source.
You'll hear: Raw kidney beans can poison you.
Holds up
The most publicized accounts report severe reactions in people eating even small amounts of raw or undercooked kidney beans.
Between July 1976 and February 1989, 50 incidents of suspected red kidney bean poisoning were reported in the UK.
Consumption of partially processed red and white kidney beans can cause GI symptoms, namely, nausea, vomiting, abdominal pain and diarrhoea.
You'll hear: Cooking destroys lectins.
Partly holds up
Boiling hard does, for most beans in the lab test (see the table above). Low heat doesn't. Chickpeas showed no activity in the main test, raw or boiled, but a second version of the test still found some after an hour's boiling. Roasting didn't stop peanut lectin reaching the blood (see the peanut card).
As lectins in pulses are thermolabile (not heat-resistant), their structure can be destroyed by heat.
Soaking (6–12 h) combined with heat treatment (≥ 100°C, > 30 min) are highly effective conditions for deactivating lectins from plant sources (e.g. PHA present in kidney beans). However, since the effect of the soaking time can vary depending on the kidney bean varieties, it is advisable to continue soaking until the beans have softened before cooking.
You'll hear: A slow cooker makes beans safe.
Doesn't hold up
Not for raw dried beans. In a lab test, heating them at 80 °C made their lectins more active:
At a heating temperature of 80 °C, there was a considerable increase in the agglutination force (clumping force), which reached a maximum after 15 minutes that was seven times higher than the initial value.
Boil them hard first, as the BfR advises (see "Slow cookers and low heat" above).
You'll hear: Canned beans are full of lectins.
Doesn't hold up
Canned beans are cooked and packaged in liquid, so they are also low in lectins.
You'll hear: Tomatoes are full of lectins.
Doesn't hold up
Tomatoes do carry some lectin, and it survived a short boil in the Danish study (see the table above). But the amounts are small:
Sugar snap peas and peas are an exception, as they contain only low amounts of lectins. They can even be eaten raw in moderation. This also applies to many other foods that contain lectins, including tomatoes, mushrooms and bananas.
You'll hear: Lectins cause leaky gut.
Not settled
Active lectins can make the gut lining more permeable (see "Do cooked lectins harm your gut?" above). Whether the amounts in cooked food do this in people hasn't been shown:
However, there is very limited research in humans on the amount of active lectins consumed in the diet and their long-term health effects.
The available epidemiological data on GI and metabolic effects are limited both in terms of volume and validity and cannot be used further for risk characterisation.
You'll hear: Lectins cause autoimmune disease.
Not settled
Because lectin proteins bind to cells for long periods of time, they can potentially cause an autoimmune response and are theorized to play a role in inflammatory conditions like rheumatoid arthritis and type 1 diabetes.
As the plant lectins hold the capacity to adversely alter the gut barrier and the microbiota, plant lectins have been implicated in the possible induction of autoimmune disorders and allergic reactions.
"Theorized" and "possible" are the key words: we found no study in people showing that lectins from cooked food cause autoimmune disease.
You'll hear: Peanut lectin gets into your blood.
Holds up
In one small study of volunteers who ate a large serving, it did, from roasted peanuts as well as raw:
Active PNA and WGA are rapidly detected in the systemic circulation after eating peanuts or wheat germ, suggesting absorption occurs in the upper GI tract, probably in the stomach.
Wang et al. (1998) assessed in healthy UK volunteers the absorption of active PNA after consumption of either raw peanuts (n = 2; 200 g; local market) or roasted salted peanuts (n = 5; 200 g) in the morning 2–3 h after a light breakfast.
PNA up to 5 μg/mL was detected in serum by immunoblot 1–4 h after ingestion in both volunteers after raw peanuts' ingestion and in four of the five volunteers after roasted peanuts' ingestion …
EFSA's account reports the lectin being found, not harm from it, and EFSA made no risk assessment for peanut lectin.
You'll hear: Lectins stop you absorbing minerals.
Partly holds up
Active lectins can, in animal and cell studies. Cooking removes most of the activity:
Animal and cell studies have found that active lectins can interfere with the absorption of minerals, especially calcium, iron, phosphorus, and zinc.
As this can reduce the absorption of nutrients in the intestines, lectins are also classified as antinutritional substances.
You'll hear: Sprouting beans and grains lowers lectins.
Partly holds up
Other processes that deactivate the compounds are sprouting grains and beans, and mechanically removing the outer hull of beans and wheat grains that contain the most lectins.
Harvard lists sprouting among processes that lower lectins, but gives no figures, so sprouting is not a substitute for boiling.
You'll hear: A lectin-free diet is healthier.
Doesn't hold up
We found no trial showing it. The BfR advises against it, and Harvard points to the population studies:
Under no circumstances should plant-based foods be avoided in favour of a low-lectin diet.
In many large population studies, lectin-containing foods like legumes, whole grains, and nuts are associated with intentional weight loss and lower rates of cardiovascular disease and type 2 diabetes.
Harvard does suggest that people with a sensitive gut may simply eat less of the foods that upset them.
The referee's verdict
- Right, for the lectin-wary side: raw and undercooked beans are genuinely dangerous; slow-cooking raw beans can make their lectins more active; active lectins can disturb the gut lining and mineral absorption in animal and cell studies; chickpeas kept some lectin activity after boiling in one version of the lab test; and peanut lectin reached volunteers' blood after raw and roasted peanuts.
- Right, for the "eat your beans" side: EFSA found no health concern from bean lectin fully deactivated by soaking and boiling; soaked and boiled beans and boiled lentils showed no lectin activity in the lab test; canned beans are low in lectins; tomatoes carry only small amounts; and legumes and whole grains are associated with lower rates of heart disease and type 2 diabetes.
- Not known: whether the lectins left in cooked food harm healthy people over years. We found very little research in people. EFSA could assess only the bean lectin, PHA; for the others it lacked the data.
Try it with your own plate
See what a plate of beans and rice gives you, and what comes with it.
Sources
- German Federal Institute for Risk Assessment (BfR). Lectins in plant-based foods: Is there a health risk? Opinion No 03/2024, 23 January 2025.Evidence: Regulatory assessmentRead: PDF text layer, in Chrome
- Harvard T.H. Chan School of Public Health. The Nutrition Source: Lectins (last reviewed January 2022).Evidence: Not applicable. A university's summary of the evidence.Read: page text, in Chrome
- Adamcová A, Laursen KH, Ballin NZ (2021). Lectin Activity in Commonly Consumed Plant-Based Foods: Calling for Method Harmonization and Risk Assessment. Foods 10(11): 2796. doi:10.3390/foods10112796Evidence: Measured in food. Lab measurements; each sample tested once.Read: PMC full text and tables, in Chrome
- Rodhouse JC, Haugh CA, Roberts D, Gilbert RJ (1990). Red kidney bean poisoning in the UK: an analysis of 50 suspected incidents between 1976 and 1989. Epidemiology and Infection 105(3): 485–491. doi:10.1017/s095026880004810xEvidence: Case report. An analysis of reported poisoning incidents.Read: PMC abstract, in Chrome
- EFSA Panel on Contaminants in the Food Chain (CONTAM) (2026). Risks for human health related to the presence of plant lectins in food. EFSA Journal 24(1): e9850. doi:10.2903/j.efsa.2026.9850Evidence: Regulatory assessmentRead: Wiley full text, in Chrome