Rice: the arsenic question, and the water you pour away

Every extra daily serving of white rice went with 11 percent more type 2 diabetes across seven cohorts and 352,384 people. Read the next line before you act on that. The link held in Asian populations, where people ate three or four servings a day, and it did not hold in Western ones eating one or two a week, and the gap between the two was itself statistically significant. Rice also has no single glycaemic index. Across varieties it runs from 48 to 93, which is the distance between a lentil and a slice of white bread, so the word on the bag tells you very little. The other half of this page is arsenic, and the one federal number in the United States is not about you. It is 100 parts per billion, it covers infant rice cereal and nothing else, and it is guidance rather than a binding limit.

The numbers, per 100 g

How much is actually in it

• Rice has no single glycaemic index: across varieties it runs from 48 to 93, which is the width of the gap between a lentil and a slice of white bread. — human postprandial studies indexed in Scopus, Medline and SciFinder to July 2014 (n = not pooled, the review reports the observed range across studies) Evidence B source
  Glycaemic index 48 to 93 against a glucose reference of 100, insulinaemic index 39 to 95; three drivers named are amylose to amylopectin ratio and cultivar, post-harvest parboiling, and consumer cooking, storage and reheating
• Cooking does not only take elements out of rice, it puts others in from the water, so raw grain and cooked grain are two different foods by composition. — 106 matched triplets of raw rice, cooked rice and urine in the Health Effects of Arsenic Longitudinal Study, Araihazar, Bangladesh (n = 106 triplets) Evidence D source
  Barium rose by a mean 241 ng/g and strontium by 679 ng/g after cooking, while caesium, selenium, molybdenum and nickel fell by 12, 19, 202 and 125 ng/g; iron, cobalt, molybdenum and vanadium in cooked rice correlated with urinary levels at Spearman rho 0.28 to 0.36
• How much arsenic is in a grain of rice is decided by where the paddy was, which is why the same variety carries different numbers in different countries. — rice commercialised from producing regions of Asia, Europe and the United States (n = a selected set of market basket and field survey datasets, not enumerated in the abstract) Evidence D source
  Geographical origin governs both bioavailability and speciation of arsenic in the grain, through groundwater irrigation, weathering of arsenic-bearing rock, mining and arsenical pesticides

What a real portion looks like

• A cup of cooked rice is not a fixed amount of food, and US labelling rules say so outright: the weight depends on how much water went in. — US food labels (n = one rulemaking) Evidence E source
  FDA kept the reference amount on the prepared basis of 1 cup while requiring the serving size to describe the product as packaged, on the reasoning that cooked weight varies with water and that a cooked basis could be manipulated for dry rice

What your body absorbs

• Which zinc salt a mill adds to fortified rice changed iron absorption by more than half, so two packets with identical labels are not the same food. — 19 iron-depleted Swiss women aged 20 to 39 with plasma ferritin at or below 16.5 µg/L and normal body mass index (n = 19) Evidence C source
  Fractional iron absorption 4.5 percent with zinc sulfate against 2.7 percent with zinc oxide (P < 0.03), a 1.6-fold difference (95% CI 1.4 to 1.9), from meals of 4 mg iron and 3 mg zinc
• Swapping rice for wheat buns as the daily staple left iron absorption unchanged at 22 percent, which argues against treating rice as a uniquely poor iron carrier. — 22 healthy urban Chinese women aged 20 to 23, allocated to a rice staple or a steamed bun staple (n = 22) Evidence C source
  Mean absorption of oral 57FeSO4 was 22.2 percent in both groups, and incorporation of infused 58FeSO4 was 93 percent against 90 percent, with no significant difference in any comparison

What cooking and storage change

• Across fifteen studies the cheapest step removed the most arsenic: rinsing beat parboiling, and both beat pressure cooking, microwaving and steaming. — rice samples, fifteen studies eligible after screening (n = 15 studies) Evidence B source
  Arsenic fell by a weighted mean difference of 0.04 mg/kg (95% CI -0.05 to -0.03, P = 0.000) after cooking, with the subgroup rank order rinsed, then parboiling, then Kateh, then high-pressure, microwave and steaming

What it does to your health

• Each extra daily serving of white rice went with an 11 percent higher rate of type 2 diabetes across seven cohorts, and the trend was linear rather than a threshold. — 352 384 adults in Asian and Western cohorts, followed 4 to 22 years (n = 13 284 incident cases among 352 384 participants) Evidence B source
  RR 1.11 (95% CI 1.08 to 1.14) per serving per day, P for linear trend < 0.001
• The rice and diabetes link held in Asian populations and did not hold in Western ones, and the difference between the two was itself statistically significant. — Chinese and Japanese cohorts eating three to four servings a day against Western cohorts eating one to two servings a week (n = seven prospective cohort analyses from four articles) Evidence B source
  Highest against lowest intake: RR 1.55 (95% CI 1.20 to 2.01) in Asian populations, RR 1.12 (0.94 to 1.33) in Western populations, P for interaction = 0.038

What people believe that the data does not support

• Cooking rice and cooling it does raise resistant starch, and the one study that tested whether that moves the glycaemic index found it did not. — varieties of rice compared for resistant starch content and glycaemic response in the cited study (n = the review names one study on rice varieties, and separately data on three heating and cooling cycles in legumes, cereals and tubers) Evidence C source
  After three heating and cooling cycles resistant starch rose from 1.86 to 3.25 percent of dry matter in cereals, and a tenfold greater resistant starch content across rice varieties produced no change in glycaemic index
• Germinated brown rice is sold on three promises in Europe and the European Commission refused all three, including the one about blood sugar. — general population of the European Union (n = three separate applications for the same substance) Evidence E source
  Claims POL-HC-8508 on blood pressure, POL-HC-8509 on weight loss and POL-HC-8510 on blood glucose all carry the status Non-authorised
• Red yeast rice is the one rice claim Europe did allow and then took back, and it was withdrawn on safety grounds rather than on doubts about the effect. — general population of the European Union (n = one entry, formerly authorised under entry IDs 1648 and 1700) Evidence E source
  Monacolins were moved to the restricted substances part of Annex III of Regulation 1925/2006, use at 3 mg or more per recommended daily portion was prohibited, and the cholesterol claim at 10 mg monacolin K per day was revoked by Commission Regulation (EU) 2024/2041

Who this works differently for

• Children who went gluten free after a coeliac diagnosis had four times more arsenic in their urine six months later, because rice is what replaces wheat. — children aged 2 to 18 with biopsy-confirmed coeliac disease, predominantly White and middle class (n = 35 of 50 eligible participants completed sample collection) Evidence C source
  Median urinary arsenic 3.3 µg/L before against 13.6 µg/L after six months on a gluten-free diet, P = 0.000004
• Rice-based baby foods exceeded the international maximum level for arsenic in three of every ten items tested worldwide. — 580 baby foods and 251 infant formulas examined across 75 studies worldwide (n = 75 studies) Evidence D source
  Arsenic median 0.048 mg/kg in rice mixes with 30 percent of items above the maximum level, against 0.165 mg/kg and 89 percent in mixed fish products
• The only federal number in the United States for arsenic in rice covers infant cereal alone, and it is guidance rather than a binding limit. — infant rice cereals sold in the United States (n = one guidance document, August 2020) Evidence E source
  Action level of 100 µg/kg, equal to 100 ppb, for inorganic arsenic in infant rice cereals

What is still unknown

• The WHO guideline on fortified rice ends with a list of what nobody knows yet, and the absorption of iron against different phytate levels is on it. — populations targeted by mass rice fortification programmes (n = one guideline, 2018) Evidence E source
  Named gaps include the bioavailability of different iron compounds, the effect of different phytate contents on iron absorption from the premix, the efficacy of fortification with nutrients other than iron, and the stability of micronutrients across cooking processes
• Europe reassessed inorganic arsenic in 2024 and reported margins of exposure below one for high consumers, meaning the calculated safety gap has closed. — European adult consumers, average and high intake (n = one assessment, drawing on the 2021 exposure assessment) Evidence E source
  Margins of exposure 2.0 to 0.4 for average adult consumers and 0.9 to 0.2 for high consumers, which the panel concluded raise a health concern

The bottom line

Rinse it. Across fifteen studies the cheapest step removed the most arsenic, ahead of parboiling, and both of those beat pressure cooking, microwaving and steaming. Then decide what to do with the water. Cooked in plenty of water you drain away, rice keeps 60 percent of its folate, 75 percent of its thiamin and 85 percent of its potassium. Cooked in water it absorbs, the same grain keeps 70, 80 and 95. Draining is the better move for arsenic and the worse move for vitamins, and nobody has run the study that tells you which trade is worth more. Cooking also adds things. In 106 matched samples from Bangladesh, barium rose by 241 nanograms per gram and strontium by 679, straight out of the water, so raw rice and cooked rice are two different foods. Two groups get far more rice than they chose. Children who went gluten free after a coeliac diagnosis had four times the arsenic in their urine six months later, 3.3 micrograms per litre before and 13.6 after, because rice is what replaces wheat. Three in ten rice-based baby foods tested worldwide sat above the international maximum. Europe reassessed inorganic arsenic in 2024 and reported margins of exposure of 2.0 to 0.4 for average adults, which means the calculated safety gap has closed. One popular claim we cannot repeat. Nobody has measured iron or zinc absorption from rice itself with isotopes, at any phytate level, so we cannot tell you that the phytate in brown rice blocks your minerals. It might. It has not been measured. Red yeast rice is a different product from the grain, and Europe allowed its cholesterol claim and then revoked it, putting monacolins under restricted substances on safety grounds rather than for any doubt about the effect.

Sources

1. rice-e-03 · systematic review
   Based on a glucose reference value of 100, the observed glycaemic index values for rice varieties ranged from 48 to 93, while the insulinaemic index ranged from 39 to 95.
   Boers et al., British Journal of Nutrition, 2015 · checked 2026-09-21
2. rice-e-06 · longitudinal observational biomonitoring
   Barium and strontium significantly increased after cooking (mean differences of 241 ng/g and 679 ng/g, respectively), while cesium, selenium, molybdenum, and nickel significantly decreased (mean differences of -12 ng/g, -19 ng/g, -202 ng/g, and -125 ng/g, respectively).
   Ecotoxicology and Environmental Safety, 2026 · checked 2026-09-21
3. rice-e-09 · meta-analytic review of composition data
   Consequently, such substantial geographical variation reflects bioavailability as well as speciation of arsenic in rice.
   Journal of Environmental Management, 2019 · checked 2026-09-21
4. rice-e-15 · final rule, agency response to comment
   Because the weight of the cooked rice depends on the amount of water used in the preparation, the amount required to make one reference amount in cooked form can vary widely.
   Food Labeling: Serving Sizes of Foods That Can Reasonably Be Consumed At One Eating Occasion, FDA, FR Doc. 2016-11865 · checked 2026-09-21
5. rice-e-10 · controlled clinical trial with stable isotopes
   In iron-depleted women, iron absorption from FePP-fortified extruded rice cofortified with ZnSO4 was 1.6-fold (95% CI: 1.4-, 1.9-fold) that of rice cofortified with ZnO.
   Journal of Nutrition, 2017 · checked 2026-09-21
6. rice-e-11 · randomised controlled trial with stable isotopes
   In all participants, the rice group, and the buns group, nonheme iron intake was 7.2±1.6, 5.9±0.6, and 8.4±1.2 mg, respectively; mean 57FeSO4 absorption rate was 22.2%±9.6%, 22.2%±10.6%, and 22.2%±8.9%, respectively
   Asia Pacific Journal of Clinical Nutrition, 2019 · checked 2026-09-21
7. rice-e-05 · systematic review and meta-analysis
   Furthermore, based on the subgroup analysis the overall rank order of cooking methods in the rice was rinsed > parboiling > Kateh > high-pressure, microwave, and steaming.
   Int J Environ Health Res, 2024 · checked 2026-09-21
8. rice-e-01 · meta-analysis of prospective cohort studies
   In the total population, the dose-response meta-analysis indicated that for each serving per day increment of white rice intake, the relative risk of type 2 diabetes was 1.11 (1.08 to 1.14) (P for linear trend<0.001).
   Hu et al., BMJ, 2012 · checked 2026-09-21
9. rice-e-02 · meta-analysis of prospective cohort studies
   The pooled relative risk was 1.55 (95% confidence interval 1.20 to 2.01) comparing the highest with the lowest category of white rice intake in Asian populations, whereas the corresponding relative risk was 1.12 (0.94 to 1.33) in Western populations (P for interaction=0.038).
   Hu et al., BMJ, 2012 · checked 2026-09-21
10. rice-e-04 · systematic review, narrative synthesis of cited trials
    After three heating/cooling cycles, the RS content of legumes, cereals and tubers increased from 4·18, 1·86 and 1·51 % to 8·16, 3·25 and 2·51 %, respectively, on a DM basis. However, a ten times greater RS content in rice varieties had no effect on the GI
    Boers et al., British Journal of Nutrition, 2015 · checked 2026-09-21
11. rice-e-12 · regulatory decision, EU health claims register
    POL-HC-8510 Vibigaba (germinated brown rice) Contributes to the maintenance of normal blood glucose levels. -/- Non-authorised
    EU Register on nutrition and health claims, European Commission, знімок 2026-09-21 · checked 2026-09-21
12. rice-e-13 · regulatory decision, EU health claims register
    POL-HC-6421 Monascus purpureous (red yeast rice) Monacolin K from red yeast rice contributes to the maintenance of normal blood cholesterol levels Following the amendment of Annex III of Regulation (EC) 1925/2006 as regards monacolins from red yeast rice, by which monacolins have been placed in Part B of Annex III, under ‘Restricted substances’, the use of monacolins from red yeast rice at levels of 3mg and more per portion of the product recommended for daily consumption is prohibited. Consequently, the health claim on monacolin K from red yeast rice related to the contribution to the maintenance of normal blood cholesterol levels, under the daily intake of 10 mg of monacolin K from red yeast rice (Entry ID: 1648, 1700), has been revoked from the Union list of permitted health claims with Commission Regulation (EU) 2024/2041. Revoked
    EU Register on nutrition and health claims, European Commission, знімок 2026-09-21, запис POL-HC-6421 · checked 2026-09-21
13. rice-e-07 · prospective cohort study
    After 6 months on a GFD, there was a significant increase in the median urinary arsenic concentration (3.3 µg/L vs 13.6 µg/L, P = 0.000004).
    American Journal of Gastroenterology, 2025 · checked 2026-09-21
14. rice-e-08 · scoping review of composition data
    For As, the highest median was found in mixed fish (0.165 mg/kg) and rice mixes (0.048 mg/kg), with 89% and 30% of items exceeding the As ML, respectively.
    Nutrition Reviews, 2026 · checked 2026-09-21
15. rice-e-16 · agency guidance
    The action level for inorganic arsenic in infant rice cereals that FDA considers achievable with the use of such practices is 100 microgram per kilogram (µg/kg), or 100 parts per billion (ppb)
    Guidance for Industry: Action Level for Inorganic Arsenic in Rice Cereals for Infants, FDA, серпень 2020 · checked 2026-09-21
16. rice-e-14 · guideline, research priorities
    During discussions in the WHO technical meeting on rice fortification, the WHO guideline development groups and the external review group highlighted the limited evidence available in some knowledge areas, meriting further research on the fortification of rice
    Guideline: Fortification of Rice with Vitamins and Minerals as a Public Health Strategy, WHO 2018 · checked 2026-09-21
17. rice-e-17 · agency risk assessment
    In adults, the margins of exposure range between 2.0 and 0.4 for average consumers and between 0.9 and 0.2 for high consumers, respectively.
    Update of the risk assessment of inorganic arsenic in food, EFSA CONTAM Panel, 2024, plain language summary · checked 2026-09-21