Heavy metals in food databaseChecked 2026.10.03
Arsenic in food
The properties of arsenic differ greatly by form. The concern is inorganic arsenic; about 70% of Japanese intake of inorganic arsenic comes from rice, followed by vegetables and seaweed (hijiki and others). Seafood is high in total arsenic (about 60% of total arsenic comes from seafood), but inorganic arsenic is only about 2%. No standard for inorganic arsenic in rice or hijiki could be found in Japan’s Food Sanitation Act. International assessments have shifted greatly (the 1988 tolerable intake was withdrawn, and there were reassessments in 2024 and 2025), and no agency has issued a tolerable intake that says “below this is safe.”
Forms
- Inorganic arsenic: The form that is abundant in rice and hijiki. IARC (2012) classifies it in Group 1, carcinogenic to humans. Assessments center on this form.
- Organic arsenic: The main form in seafood is arsenobetaine, and the main form in seaweed is arsenosugar. Its toxicity is said to be low, but no official document asserts that “organic arsenic is harmless.” MAFF writes both that the degree of harm is generally said to be small and that what effects it has in the human body is not well understood at present.
- Total arsenic: The sum of inorganic and organic. Japanese intake is about 200 μg/day, of which about 9% is inorganic arsenic.
- Pay attention to the state: Dried hijiki versus soaked (rehydrated) hijiki, and brown rice versus polished (white) rice, differ by orders of magnitude. This database always states both the form (total arsenic or inorganic arsenic) and the state (dried, soaked, or raw).
Foods with arsenic data
| Food | Form / state | Mean | Range / max |
|---|---|---|---|
| Hijiki | Total arsenic, dried | 93 (median 92) | 28–160 |
| Hijiki | Inorganic arsenic, dried | 67 (median 68) | 4.5–130 |
| Kombu | Total arsenic, dried | 53 (median 51) | 25–110 |
| Wakame | Total arsenic, dried | 33 (median 33) | 15–52 |
| Nori | Total arsenic, dried | 25 (median 23) | 13–51 |
| Hijiki | Total arsenic, after soaking (the item name in MAFF’s table is “soaked product”) | 6.0 (median 5.5) | 2.1–20 |
| Hijiki | Inorganic arsenic, after soaking (the item name in MAFF’s table is “soaked product”) | 3.6 (median 3.2) | <0.5–17 |
| Baby food | Total arsenic, dry type | 0.65 (median 0.021) | max 8.6 |
| Brown rice | Total arsenic, brown rice | 0.23 (median 0.21) | max 0.80 |
| Brown rice | Inorganic arsenic, brown rice | 0.21 (median 0.20) | max 0.59 |
| Brown rice | Inorganic arsenic, brown rice | 0.17 (median 0.16) | max 0.60 |
| Brown rice | Inorganic arsenic, brown rice | 0.15 (median 0.14) | max 0.37 |
| Brown rice | Inorganic arsenic, brown rice | 0.15 (median 0.14) | max 0.38 |
| Brown rice | Inorganic arsenic, brown rice | 0.15 (median 0.15) | max 0.37 |
| Polished rice | Total arsenic, polished rice | 0.14 (median 0.13) | max 0.44 |
| Brown rice | Inorganic arsenic, brown rice | 0.13 (median 0.12) | max 0.37 |
| Polished rice | Inorganic arsenic, polished rice | 0.12 (median 0.12) | max 0.26 |
| Polished rice | Inorganic arsenic, polished rice (paddy soil survey, etc.) | 0.10 (median 0.09) | max 0.35 |
| Polished rice | Inorganic arsenic, polished rice | 0.098 (median 0.10) | max 0.25 |
| Polished rice | Inorganic arsenic, polished rice | 0.092 (median 0.08) | max 0.26 |
| Polished rice | Inorganic arsenic, polished rice | 0.09 (median 0.09) | max 0.23 |
| Spinach | Total arsenic | 0.01 | max 0.12 |
| Taro and potatoes | Total arsenic, taro (unpeeled) | 0.01 | max 0.04 |
| Wheat | Total arsenic, wheat (domestic) | 0.009 | max 0.04 |
| Soybeans | Total arsenic, soybeans (domestic) | 0.008 | max 0.04 |
| Taro and potatoes | Total arsenic, taro (peeled) | 0.006 | — |
| Mackerel | Mackerel, total arsenic (edible portion) | (no mean given) | 1.5–2.7 |
| Mackerel | Chub mackerel, total arsenic (muscle, dry weight) | 5.4 | — |
| Mackerel | Chub mackerel, inorganic arsenic (muscle, dry weight) | 0.00 (not detected) | — |
| Sardines, horse mackerel, and Pacific saury | Sardine, total arsenic (edible portion) | (no mean given) | 2.5–3.5 |
| Sardines, horse mackerel, and Pacific saury | Horse mackerel, total arsenic (edible portion) | (no mean given) | 1.9–7.0 |
| Sardines, horse mackerel, and Pacific saury | Japanese sardine, total arsenic (muscle, dry weight) | 17.3 | — |
| Sardines, horse mackerel, and Pacific saury | Japanese sardine, inorganic arsenic (muscle, dry weight) | 0.28 | — |
| Sardines, horse mackerel, and Pacific saury | Japanese horse mackerel, total arsenic (muscle, dry weight) | 25.6 | — |
| Sardines, horse mackerel, and Pacific saury | Japanese horse mackerel, inorganic arsenic (muscle, dry weight) | 0.06 | — |
| Sardines, horse mackerel, and Pacific saury | Pacific saury, total arsenic (muscle, dry weight) | 5.5 | — |
| Sardines, horse mackerel, and Pacific saury | Pacific saury, inorganic arsenic (muscle, dry weight) | 0.22 | — |
| Kombu | Inorganic arsenic, dried | <0.5 (all 100 samples below the limit of quantification) | — |
| Wakame | Inorganic arsenic, dried | <0.5 (all 100 samples below the limit of quantification) | — |
| Nori | Inorganic arsenic, dried | <0.5 (all 100 samples below the limit of quantification) | — |
| Green tea | Total arsenic, tea leaves (dried) | Not calculated (119 samples below the limit of quantification) | <0.01–0.13 |
| Baby food | Inorganic arsenic, dry type | 0.043–0.047 | max 0.80 |
| Baby food | Inorganic arsenic, wet type | 0.006–0.013 | max 0.17 |
| Baby food | Inorganic arsenic, snacks for infants and toddlers | 0.022–0.027 | max 0.095 |
| Baby food | Total arsenic, wet type | 0.043–0.047 (median 0.021) | max 0.44 |
| Infant formula | Inorganic arsenic, formula (powder, per powder) | 0.007–0.013 | max 0.047 |
| Infant formula | Inorganic arsenic, formula (powder, estimate after mixing, mg/L) | 0.0010–0.0017 | — |
| Infant formula | Inorganic arsenic, formula (liquid) | 0.002–0.004 | max 0.004 |
| Infant formula | Total arsenic, formula (powder, per powder) | 0.014–0.018 (median 0.011) | max 0.069 |
Sorted by mean, highest first. Values in different forms (total vs inorganic, dried vs soaked) cannot be compared directly — check each food page.
Effects on the body
- Most of the epidemiological evidence comes from regions where drinking water was contaminated with arsenic (Taiwan, Chile, Bangladesh, West Bengal in India, and Inner Mongolia in China). Both the FSCJ (differences in nutritional status, and mixing of water and diet) and EFSA (the epidemiology of skin lesions comes from low- and middle-income regions, so it is hard to apply to Europe) write about the uncertainty in applying this evidence to low dietary levels.
- IARC (2012) classified arsenic and inorganic arsenic compounds in Group 1 (carcinogenic to humans). The basis is epidemiology showing that arsenic in drinking water increases bladder, lung, and skin cancers.
- Studies of the Japanese diet. In the JPHC cohort (90,378 people, 11 years, 7,002 cancer cases), dietary arsenic was not associated with cancer overall, but there was a tendency for association with lung cancer in men, which became stronger in men who currently smoke (hazard ratio 1.36, 95% confidence interval 1.09–1.70 in the highest inorganic arsenic group, Sawada et al. 2013). Intake was estimated from a dietary questionnaire and the effect of smoking is mixed in, so this result alone does not show that arsenic causes lung cancer.
- The FSCJ (December 2013) did not set a tolerable daily intake. It concluded that there is no data showing health effects from a normal diet, that it does not consider the current situation problematic, and that people should eat a balanced diet without favoring particular foods. On the other hand, it also wrote that the estimated intake of inorganic arsenic is “not far from” the no-effect levels (NOAEL and BMDL) calculated in the assessment, and that some people with high intake may exceed them.
- At high doses, arsenic causes acute poisoning. Examples include the 1955 Morinaga arsenic milk poisoning and the 1998 Wakayama arsenic curry poisoning. The estimated intake of survivors of the Wakayama incident averaged 53 mg, while inorganic arsenic in the Japanese diet is about 18 μg a day, so they differ by orders of magnitude.
Tolerable intake (the amount thought to cause no harm even if taken for a lifetime)
| Body | Value | Note |
|---|---|---|
| JECFA (1988) | PTWI 15 μg/kg body weight/week (about 2.1 μg/kg/day) | Withdrawn in 2010. It was close to the BMDL0.5 for lung cancer (3.0 μg/kg/day), and was judged not to protect health |
| JECFA (2010, 72nd meeting) | BMDL0.5 3.0 (2–7) μg/kg body weight/day | Lung cancer. Bladder cancer is 5.2 and skin lesions 5.4. No new tolerable intake was issued |
| JECFA (2025, 101st meeting) | POD 0.3 μg/kg body weight/day | BMDL0.5 for ischemic heart disease. This value was confirmed in a published summary |
| EFSA (2009) | BMDL01 0.3–8 μg/kg body weight/day | Given as a range rather than a single value. EFSA judged a tolerable intake inappropriate and evaluated by the margin of exposure (MOE) against intake |
| EFSA (January 2024) | BMDL05 0.06 μg/kg body weight/day | Skin cancer. Mean exposure of European adults is 0.03–0.15 and the 95th percentile is 0.07–0.33; the MOE is 2–0.4 at the mean and 0.9–0.2 at the 95th percentile. EFSA concluded it “raises a concern.” It has not defined an MOE threshold for “low concern” |
| FSCJ (2013) | Not set | Reasons: deriving a threshold for the Japanese diet from epidemiology in drinking-water-contaminated areas gives large errors, and knowledge of the mechanism of carcinogenesis is insufficient |
| U.S. EPA (1991) | RfD 0.3 μg/kg body weight/day | The value used for comparison in the National Institute of Health Sciences report (2024) |
| Tap water (water quality standard, WHO) | 0.01 mg/L (10 μg/L) | Japan’s water quality standard and the WHO guideline value are the same. This is a drinking-water standard, not a tolerable intake from the diet |
How much people in Japan take in
- Inorganic arsenic intake was 18.3 μg/day by the market basket method (average of fiscal 2014–17). The breakdown was rice 12.2 (about 67%), vegetables and seaweed 2.9, seasonings 1.1, and seafood 0.4. In the 2024 total diet study it was 21.2 μg/day (0.39 μg/kg/day per body weight).
- Total arsenic is about 200 μg/day (market basket 205, duplicate diet 199), of which seafood is 121.0 (59.0%), vegetables and seaweed 62.2 (30.3%), and rice 12.6 (6.1%). The leading source switches from seafood for total arsenic to rice for inorganic arsenic.
- By the duplicate diet method (samples from a Ministry of the Environment survey, fiscal 2006–10, 319 meals), inorganic arsenic averaged 18.6 (2.18–161) μg/day, and the 95th percentile per body weight was 0.754 μg/kg/day. Estimates vary by survey, from 6.5 to 33.7 μg/day (Mohri 1990, Yamauchi 1992, Oguri 2012, Yoshinaga 2025, and others).
- Trend over time: Total arsenic rose from a low of 145 μg/day in 2007 (an approximation read from a graph) to 322 in 2021 and 219 μg/day in 2024. Seafood consumption has fallen and seaweed consumption is almost constant, so the report speculates that concentrations in food may be increasing. Inorganic arsenic has been almost constant, between 14 and 23 μg/day, since 2014.
- Placed against the tolerable levels: In 2024, inorganic arsenic had a hazard ratio of 1.28–1.80 against the EPA RfD and the EFSA 2005 provisional TWI, among the yardsticks the report chose, and it was the only one of the four metals above 1. Against the JECFA BMDL0.5 (3.0 μg/kg/day) it is not a ratio but a margin of exposure (MOE) of 7.8. However, the FSCJ in Japan has not set a tolerable level.
- How it looks depends on the yardstick. Dividing the average inorganic arsenic intake of 18.3 μg/day by a body weight of 60 kg gives 0.305 μg/kg/day. This is about 15% of the withdrawn JECFA PTWI (about 2.1), about one-tenth of the JECFA 2010 BMDL0.5 (3.0), almost the same as the JECFA 2025 POD (0.3), and about 5 times the EFSA 2024 reference point (0.06). This is a way of laying the comparisons out, not a line between safe and dangerous.
History
- 1955: The Morinaga arsenic milk poisoning. Disodium phosphate added to powdered milk at the Tokushima plant was contaminated with about 10% inorganic arsenic (pentavalent). About 12,000 infants were poisoned and 133 are said to have died (FSCJ assessment).
- December 28, 1959: Prompted by this incident, the standards and specifications for foods and additives (Ministry of Health and Welfare Notification No. 370) were established, and the regulation of arsenic began.
- 1998: The Wakayama arsenic curry poisoning (diarsenic trioxide). 67 people suffered acute poisoning and 4 died.
- 2003: In Ibaraki Prefecture, well water was contaminated by illegally dumped waste containing diphenylarsinic acid, and residents suffered health damage (MAFF).
- July 2004: The UK Food Standards Agency advised people not to eat hijiki, and the MHLW issued a Q&A two days later.
- 2010: JECFA withdrew the PTWI (15 μg/kg per week). 2012: IARC classified arsenic in Group 1. December 2013: The FSCJ completed its assessment.
- 2014 and 2016: Codex adopted 0.2 mg/kg for polished rice (2014) and 0.35 mg/kg for brown rice (2016). In July 2017 it adopted a code of practice for preventing and reducing arsenic contamination in rice.
- 2020: The U.S. FDA set an action level of 100 μg/kg for inorganic arsenic in infant rice cereal (final guidance).
- January 2024: EFSA set its reference point at a BMDL05 of 0.06 μg/kg/day. 2025: JECFA reassessed and derived a POD of 0.3 μg/kg/day from ischemic heart disease.
Good to know
- No standards for inorganic arsenic in rice, hijiki, or seafood could be found in Japan (this is as far as the tables in the FSCJ assessment and the MAFF materials were read; the legal texts were not checked one by one). Because we cannot write “exceeded the standard,” when applying overseas yardsticks we write “if another country’s standard is applied.”
- Countries’ views on hijiki are divided. The UK, Canada, Hong Kong, Australia, and New Zealand say to avoid it, while Japan’s MHLW, MAFF, and FSCJ say “it is fine if you eat a balanced diet; use cooking methods that reduce it.” We have found no document in which the authorities of these countries compare their reasons and explain them.
- The calculation in the MHLW Q&A (2004) was based on the JECFA PTWI (15 μg/kg per week) that was later withdrawn. It is a calculation as of 2004 and should not be cited as a current assessment.
- In rice cultivation there is a trade-off between arsenic and cadmium. Draining the paddy water after heading reduces arsenic but increases cadmium. In June 2024, MAFF drew up guidelines to reduce both.
- We could not find any public survey of inorganic arsenic in processed rice products in Japan (rice crackers, rice flour, rice bran, and rice-bran pickles). The EU has standards such as 0.30 mg/kg for rice crackers. We have also found no public data on seaweeds in the same family as hijiki (Sargassaceae), such as mozuku and arame.
- Flat statements such as “hijiki is high in arsenic, so don’t eat it” or “brown rice is high in arsenic, so avoid it” do not match the views of MAFF and the FSCJ. On the other hand, statements that flatly say “arsenic is no problem” also do not match the assessments in which EFSA 2024 and JECFA 2025 found a “possible concern.” We set both side by side and leave the judgment to the reader.
- EFSA’s exposure figures for European adults cannot be applied as they are to Japanese people, because the foods (rice and seaweed) are different.
Related explainers
- Should You Avoid Hijiki? Inorganic Arsenic Numbers, Ways to Reduce It, and Views by Country
- Arsenic and Cadmium in Brown and White Rice: Is Eating It Every Day OK? Measurements of Japanese Rice and How to Reduce Them
- Heavy Metals in Baby Food, Infant Food, and Powdered Formula: MAFF’s Survey and Standards Abroad
- Can Home Cooking Reduce Heavy Metals? Figures Shown for Hijiki, Rice, Innards, Tap Water, and Canned Food
- Why Do Heavy Metal Standards Differ Between Japan and Other Countries? Japanese, Codex, and EU Standards and Tolerable Intakes Compared
- How Much Heavy Metal Do Japanese People Take In? Fifty Years of Lead, Cadmium, Mercury, and Arsenic Intake
Sources
- MAFF, “Survey of arsenic contained in food” (updated December 19, 2025)
- MAFF, arsenic intake from rice and comparison with other countries
- MAFF, arsenic in hijiki (home-cooking test)
- FSCJ, food health effects assessment, “Arsenic” (December 2013)
- EFSA CONTAM Panel. EFSA J 2024;22(1):e8488 (inorganic arsenic)
- JECFA database (arsenic)
- MHLW Health and Labour Sciences Research Grants, total diet study, fiscal 2024 (National Institute of Health Sciences)
- Sawada N, et al. Cancer Causes Control 2013 (JPHC cohort and arsenic)（PMID 23666560）
Last checked 2026.10.03 (against the original government documents and papers, collected and cross-checked in Gemini Notebook).
Heavy metals occur naturally in soil and the sea, and every food contains small amounts. This page lists facts from government surveys, limits and assessments and from published papers; it does not judge any food as good or bad. Where there is official advice — such as on eating fish during pregnancy — follow it.