Beans: what half a cup a day does to cholesterol

Beans are one of the few ordinary foods with a dose-response curve behind them. Across 38 randomized trials with 2095 adults, a median of 4.6 oz (130 g) of whole pulses a day lowered LDL cholesterol by 5.4 mg/dL (0.14 mmol/L), which is about 5 mg/dL, and those trials ran a median of six weeks. For beans on their own the effect grew with the portion, by 0.25 mmol/L of LDL for every half cup a day, though the confidence interval on that slope runs from 0.48 down to 0.02, so its weak end is close to no effect at all. That line was tested only as far as one cup a day. Past a cup nobody has looked. Blood sugar, the thing beans are more often sold on, held up worse. Pooling 23 trials of common beans, the glucose peak at 60 minutes came out 2.01 mg/dL lower. The authors call that difference not significant, even though the interval they print with it, 4.6 to 0.63, does not contain zero, so how big it is cannot be read off their paper. The trials also disagreed with each other so badly (heterogeneity 98 percent) that the pooled number carries almost nothing. Iron is the third thing beans are known for. Measured with stable isotopes in 29 healthy young adults in Brazil who were not anemic, a single 3.5 oz (100 g) meal of cooked common beans gave up about 0.41 percent of its iron, and beans bred to contain more iron gave up the same fraction, 0.407 percent against 0.409. If you cook dry kidney beans, the FDA says to soak them for at least 5 hours and boil them in fresh water for at least 30 minutes, because as few as four or five raw beans can trigger symptoms.

The numbers, per 3.5 oz (100 g)

As people eat it, per 3.5 oz (100 g)

What a real portion looks like

• For beans the effect grew with the portion, by 0.25 mmol/L of LDL for every half cup a day up to one cup. — adults in the same 38 trials, dose-response analysis restricted to beans (n = 2095) Evidence A source
  -0.25 mmol/L LDL per 0.5 cup a day (95% CI -0.48 to -0.02) and -17 mg/dL (-0.45 mmol/L) non-HDL per 0.5 cup (95% CI -0.71 to -0.18), linear up to 1 cup a day

What your body absorbs

• Only about 0.41 percent of the iron in a 3.5 oz (100 g) meal of cooked common beans was absorbed, and beans bred for higher iron did no better at 0.407 percent. — 29 young adult volunteers in Brazil, 13 on ordinary beans and 16 on beans bred for iron, stable isotope method (n = 29) Evidence B source
  Iron absorption 0.409 percent from ordinary beans against 0.407 percent from biofortified beans, no difference between groups
• Across nine efficacy studies, legume varieties bred for lower phytate or higher minerals raised iron intake at least 1.5-fold and lifted hemoglobin by up to 3.0 g/L. — participants of nine intervention studies in low and middle income countries (n = 9 studies) Evidence C source
  Iron intake up at least 1.5-fold, hemoglobin up by as much as 3.0 g/L, no efficacy data for zinc status at all

What cooking and storage change

• Adzuki beans landed in the low glycemic index band of 21 to 51 while rice and the other test foods sat at 83 to 109, and pre-soaking pushed starch toward the rapidly digestible form. — 10 healthy adults, crossover, six starchy foods each cooked for two durations (n = 10) Evidence B source
  Glycemic index 21 to 51 for adzuki bean and lotus seed against 83 to 109 for the other foods; hydrolysis index 37.7 to 61.1 percent
• The FDA says slow cookers should not be used for kidney beans or dishes containing them, because slow-cooked casseroles often reached only 167 °F (75°C) inside, too low to destroy the lectin. Raw kidney beans carry 20,000 to 70,000 hemagglutinating units and fully cooked beans 200 to 400. — consumers cooking kidney beans or dishes that contain them in slow cookers Evidence E source
  Slow-cooked casseroles often reached only 167 °F (75°C) or less inside. Raw kidney beans 20,000 to 70,000 hau, fully cooked 200 to 400 hau, white kidney beans about one-third of red

What it does to your health

• Whole pulses at a median dose of 130 g a day lowered LDL cholesterol by 5.4 mg/dL (0.14 mmol/L) across 38 randomized trials with 2095 participants. — 2095 adults in 38 trials lasting at least 3 weeks, median 6 weeks (n = 2095) Evidence A source
  LDL cholesterol -5.4 mg/dL (95% CI -7.3 to -3.1; metric: -0.14 mmol/L, -0.19 to -0.08), non-HDL -8.5 mg/dL (-0.22 mmol/L), apoB -0.08 g/L, HDL -1.2 mg/dL (-0.03 mmol/L)

Safety, limits and interactions

• The FDA says dry kidney beans should be soaked for at least 5 hours, the water poured away, and the beans boiled in fresh water for at least 30 minutes. It notes that as few as four or five raw beans can trigger symptoms. — consumers cooking dry kidney beans (Phaseolus vulgaris), all ages Evidence E source
  Soak at least 5 hours, discard the water, boil in fresh water at least 30 minutes. As few as four or five raw beans can trigger symptoms

What people believe that the data does not support

• A meta-analysis of 23 trials on common beans found a 2.01 mg/dL lower postprandial glucose peak at 60 minutes, a difference that was not significant and sat on heterogeneity of 98 percent. — human participants of 23 clinical and randomized trials selected from 340 screened articles (n = 23 trials) Evidence A source
  Mean difference -2.01 (95% CI -4.6 to -0.63) for the 60 minute glucose peak, not significant, I2 = 98 percent
• The lectins of Phaseolus vulgaris act as anti-nutritional factors that can upset the gut, and heating is one of the treatments that decides whether they stay active. — laboratory and processing studies of Phaseolus vulgaris lectins, no clinical trial Evidence D source
  Lectins described as toxic allergens and hemagglutinins whose activity is eliminated or changed by proteolysis, heating, gamma radiation and high hydrostatic pressure

Who this works differently for

• The FDA says all people, regardless of age or gender, appear equally susceptible to kidney bean lectin poisoning, that severity depends on the dose eaten, and that the attack rate in seven UK outbreaks was 100%. — people eating raw or undercooked kidney beans Evidence E source
  No group named as more susceptible. Severity related to dose, attack rate 100% in seven UK outbreaks
• In 55 Italian children with legume allergy, where common bean was one of the legumes studied, 12.7% had anaphylaxis at their first reaction, three in four reacted to more than one legume, and pea and lentil were the most frequent triggers. — 55 children under 18 allergic to at least one of pea, lentil, chickpea, common bean or soy (n = 55) Evidence C source
  Anaphylaxis at onset 12.7%. Pea 70.9% and lentil 69.1% most prevalent. Multi-legume allergy 74.5%. Peanut co-allergy 25.5%

The bottom line

Cook dry beans properly and do not shortcut it. Raw or undercooked kidney beans carry a lectin, phytohaemagglutinin, and the FDA says as few as four or five raw beans can trigger symptoms. Its advice is to soak dry kidney beans for at least 5 hours, pour the water away and boil them in fresh water for at least 30 minutes. It warns against slow cookers for kidney beans or dishes that contain them, because slow-cooked casseroles often reached only 167 °F (75°C) inside, too low to destroy the lectin. Raw kidney beans carry 20,000 to 70,000 hemagglutinating units and fully cooked beans 200 to 400. The FDA names no group as more at risk. Everyone appears equally susceptible, severity depends on the dose eaten, and in seven UK outbreaks everyone who ate the beans fell ill. Beans can also be allergens. In 55 Italian children with legume allergy, where common bean was one of the legumes studied, 12.7 percent had anaphylaxis at their first reaction and three in four reacted to more than one legume, with pea and lentil the most frequent triggers. That was one clinic, and it gives no separate figure for common bean. We hold no card on beans and medicines, and none on pregnancy or kidney disease. If you are in one of those groups, this page does not answer you, and we would rather say so than fill the space. Soaking also changes the starch. In ten healthy adults, pre-soaking pushed bean starch toward the rapidly digestible form and away from resistant starch. That is a finding about starch, and it is no reason to skip the soak. The low glycemic index in that same trial, 21 to 51 against 83 to 109 for rice and the other test foods, was measured on adzuki beans, so it is not a number about the kidney beans in your cupboard. On iron the one hopeful line comes from nine intervention studies in low and middle income countries, where legumes bred for lower phytate or higher minerals raised iron intake at least 1.5-fold and lifted hemoglobin by as much as 3.0 g/L. Those studies ran where iron deficiency is common, which is not most kitchens reading this. For zinc the same review found no efficacy data at all. Two more holes worth naming out loud. The USDA retention table we use for cooking losses carries neither vitamin K nor fiber, so for those nutrients we have trials or we have nothing. And the data table on this page runs to 138 records, from plain boiled black beans to baked beans canned with pork and franks. The trials above were run on the first kind of food, and the numbers do not transfer to the second.

Sources

1. bns-r3-02 · meta-analysis of randomized controlled trials
   A linear inverse relationship was shown for beans up to 1 cup/d for low-density lipoprotein-cholesterol (coefficient, -0.25 mmol/L/0.5 cup [95% CI, -0.48 to -0.02]) and non-high-density lipoprotein-cholesterol (-0.45 mmol/L/0.5 cup [95% CI, -0.71 to -0.18]).
   J Am Heart Assoc, 2026 · checked 2026-09-23
2. bns-r3-04 · randomized controlled trial
   Mean iron absorption from the meal with common beans was 0.409% (±0.040%) and mean iron incorporation from the meal with target beans for biofortification 0.407% (±0.038%) and did not differ between the groups.
   Clin Nutr ESPEN, 2018 · checked 2026-09-23
3. bns-r3-05 · systematic review of intervention studies
   Consumption of improved legumes resulted in a ≥1.5-fold increase in iron intake...The prevalence of iron deficiency (ID) decreases with consumption of improved legumes, with increases of ≤3.0 g/L in Hb concentrations.
   Adv Nutr, 2020 · checked 2026-09-23
4. bns-r3-06 · randomized crossover trial
   Both the LS and AB meals achieved low glycemic index (GI 21⁻51), while the other starchy foods failed to show significant difference with rice (GI 83⁻109)...The in vitro tests indicated that pre-soaking resulted in high rapidly digestible starch (RDS) and low resistant starch (RS).
   Nutrients, 2019 · checked 2026-09-23
5. hf2-bns-2 · agency guidance
   Raw kidney beans contain from 20,000 to 70,000 hau, while fully cooked beans contain from 200 to 400 hau. White kidney beans, another variety of Phaseolus vulgaris, contain about one-third the amount of toxin as the red variety ... Slow cookers should not be used to cook these beans or dishes that contain them. Studies of casseroles cooked in slow cookers revealed that the food often reached internal temperatures of only 75°C or less, which is inadequate for destruction of the toxin.
   US FDA, Bad Bug Book, 2nd edition (2012), chapter Phytohaemagglutinin (kidney bean lectin) · checked 2026-10-08
6. bns-r3-01 · meta-analysis of randomized controlled trials
   Thirty-eight trials (52 trial comparisons, n=2095) with a median of 6 weeks and dose of 130 g/d (0.5-0.67 cup/d) showed that whole dietary pulses decreased low-density lipoprotein-cholesterol (mean difference, -0.14 mmol/L [95% CI, -0.19 to -0.08])
   J Am Heart Assoc, 2026 · checked 2026-09-23
7. hf2-bns-1 · agency guidance
   Studies done by British scientists suggest that beans should be soaked in water for at least 5 hours, the water poured away, and the beans boiled in fresh water for at least 30 minutes. ... Toxic dose: As few as four or five raw beans can trigger symptoms. ... Consumers should boil the beans for at least 30 minutes to ensure that the product reaches sufficient temperature, for a sufficient amount of time, to completely destroy the toxin.
   US FDA, Bad Bug Book, 2nd edition (2012), chapter Phytohaemagglutinin (kidney bean lectin) · checked 2026-10-08
8. bns-r3-03 · meta-analysis of clinical and randomized controlled trials
   However, diabetes studies indicated that the postprandial glucose level at the peak of 60 min for common bean consumers was low (mean difference = -2.01; 95% CI [-4.6, -0.63]), but the difference between the treated and control was not significant, and there was a high level of heterogeneity among studies (I2 = 98%).
   Nutrients, 2021 · checked 2026-09-23
9. bns-r3-07 · systematic review of in vitro and processing studies
   Studies have shown that lectins are anti-nutritional factors that can cause intestinal disorders. Owing to their ability to act as toxic allergens and hemagglutinins, the Phaseolus vulgaris lectins are of grave concern for human health and safety.
   Crit Rev Food Sci Nutr, 2018 · checked 2026-09-23
10. hf2-bns-3 · agency guidance
    All people, regardless of age or gender, appear to be equally susceptible; the severity is related to the dose ingested. In the seven outbreaks mentioned below, the attack rate was 100%.
    US FDA, Bad Bug Book, 2nd edition (2012), chapter Phytohaemagglutinin (kidney bean lectin) · checked 2026-10-08
11. hf2-bns-4 · retrospective cohort
    This retrospective single-center study (January 2022-January 2024) included children (<18 years) allergic to ≥1 index legume (pea, lentil, chickpea, common bean, or soy). ... Median age at first reaction was 18 months; anaphylaxis occurred at onset in 12.7%, most frequently triggered by pea. Pea (70.9%) and lentil (69.1%) were the most prevalent allergies, with pea causing 50% of index-legume anaphylaxis. Multi-legume allergy predominated (74.5%)
    Nutrients, 2026 · checked 2026-10-08