There is a piece of folk pharmacology almost everyone has absorbed: fat-soluble vitamins build up and can hurt you, while water-soluble ones pass through, so an extra B vitamin is at worst a waste of money. The belief is tidy, widely repeated, and for at least one vitamin it is simply wrong.
In August 1983, the New England Journal of Medicine published a paper by Herbert Schaumburg and colleagues describing seven adults with ataxia, meaning a loss of control over bodily movement, and severe damage to the sensory nervous system after taking high daily doses of vitamin B6. Four were severely disabled. All improved after stopping. The authors called it a new megavitamin syndrome, which tells you how unexpected it was, and urged that safe guidelines be established for a vitamin they described as widely abused.
For forty years, that series was taken to describe a megadose phenomenon, and the reference figures reflect it. The NIH Office of Dietary Supplements still describes the classic picture as chronic administration of 1 to 6 grams of oral pyridoxine a day for 12 to 40 months, which is ten to sixty times the United States upper limit, sustained for years.
Then the picture moved. In 2025, an Australian regulator published adverse event data that does not look like 1983 at all.
What Australia’s regulator reported, and when
On 25 November 2025, Australia’s Therapeutic Goods Administration issued a notice of final decision to amend that country’s Poisons Standard for pyridoxine, pyridoxal and pyridoxamine, the three forms of vitamin B6. This is a non-US regulator under non-US law, so nothing in it changes an American figure. What it supplies is evidence.
As of 31 October 2025, the agency’s adverse event database held 250 reports of peripheral neuropathy involving products containing vitamin B6. The timing is worth pausing on. A majority, 208 or 83 percent, were reported since 1 January 2023, and 130 during 2025 alone up to the end of October.
The doses are not gram-level. Citing New South Wales poisons information centre data, the notice states that patients reported adverse effects after taking daily pyridoxine doses between 25 milligrams and 400 milligrams, for as little as two weeks and as long as ten years. Both ends matter, because this is not only a story about brief exposure.
From submissions made during the consultation, the notice reports that toxicity occurred at doses below Australia and New Zealand’s 50 milligram upper level of intake in 31 individuals, 12 of whom had been diagnosed or tested for blood B6 levels. More than 100 individuals, 103 of them, reported what the agency calls “severe, sometimes permanent, health effects.” The official conducting the review concluded there is sufficient evidence that peripheral neuropathy may occur in some individuals at a B6 intake of 50 milligrams a day or even lower.
Australia’s response is a restriction. From 1 June 2027, oral products containing more than 50 milligrams but not more than 200 milligrams per recommended daily dose become pharmacist-only medicines, and those above 200 milligrams remain prescription-only. Products at 50 milligrams or less stay on general sale. That date is still in the future, so the detail is worth re-checking.
Three authorities, three different ceilings
Here the subject becomes genuinely unsettled rather than merely alarming. Three authoritative bodies have set three different upper limits for vitamin B6, differing by a factor of eight.
- United states: 100 milligrams a day for adults 19 and over. ODS states these limits apply to intake from both food and supplements, and do not apply to people receiving B6 as medical treatment under a physician’s care.
- Australia and new Zealand: 50 milligrams a day, the figure the Australian notice cites throughout.
- European food safety authority: 12 milligrams a day for all adults, set in 2023 on the basis of systematic reviews examining vitamin B6 and peripheral neuropathy, as reported on the ODS fact sheet.
Same nutrient. Same harm. Eight-fold spread. No resolution is offered here, because none exists in the sources.
The US figure’s derivation explains part of the gap. ODS says the Food and Nutrition Board took the doses used in the studies it relied on and halved them, citing limitations in the long-term data. A limit built that way is a judgment call about uncertainty, and different bodies reviewing different evidence have made it differently.
The underlying evidence is also internally mixed, and the ODS page says so on one screen. The Food and Nutrition Board noted that several reports show sensory neuropathy at doses lower than 500 milligrams a day. The same page reports that studies of patients treated with an average dose of 200 milligrams a day for up to five years found no evidence of the effect. Both come from the agency.
Whether the damage reverses is also unsettled. The 1983 series reported that all seven improved after withdrawal. ODS says symptoms usually stop if someone discontinues the supplements as soon as neurologic symptoms appear, which is a conditional sentence rather than a guarantee. The Australian notice describes effects that can be permanent. The honest composite: improvement is common when exposure stops early, and permanent harm is documented.
An upper limit is not a poisoning line
Article 323 covers what a tolerable upper intake level is and what it includes. What belongs here is narrower: a limit is a conservative margin, and the distance between a limit and actual harm runs in both directions.
In one direction the margin is generous. ODS gives the adult limit for vitamin D as 100 micrograms, or 4,000 international units, then states that signs and symptoms of toxicity are unlikely at daily intakes below 250 micrograms, which is 10,000 international units. That is a 2.5-fold gap between the ceiling and where ODS expects trouble.
Niacin goes further. Its adult limit is 35 milligrams a day for supplemental niacin, and ODS says the limits are based on the levels associated with skin flushing. Flushing, with burning, tingling and itching, typically starts at 30 to 50 milligrams of nicotinic acid, often within half an hour. ODS calls these effects exactly what they are: “they are considered an unpleasant, rather than a toxic, side effect.” A limit set on an endpoint the agency calls non-toxic cannot be read as a poisoning threshold.
Niacin shows the number alone is incomplete in two further ways. ODS states the limit does not apply to people receiving supplemental niacin under medical supervision, and doses of 1,000 to 3,000 milligrams a day are legitimately prescribed for abnormal blood lipids, up to 86 times the limit. And formulation changes risk at an identical dose: ODS reports liver toxicity is more likely with extended-release forms of nicotinic acid. A milligram figure without the form attached does not describe a risk.
Now the opposite direction, the half that gets left out. Harm has appeared well below some of these limits.
ODS gives the limit for supplemental vitamin E as 1,000 milligrams, which it equates with 1,100 international units of the synthetic form. Yet the large SELECT trial used 400 international units a day of synthetic vitamin E, roughly 36 percent of that ceiling, and follow-up found a 17 percent increase in prostate cancer risk among the men who took it. ODS also reports trials finding increased hemorrhagic stroke risk at high doses.
Vitamin D supplies the other example. In a trial of 36,282 postmenopausal women that ODS summarizes, kidney stone risk rose 17 percent across seven years among participants whose calcium ran high, near 2,100 milligrams daily, while their vitamin D sat at only 19 micrograms, equivalent to 765 international units. That second figure is under a fifth of the ceiling named earlier in this section.
So the rule is not that limits sit too high or too low. Exceeding one does not mean being poisoned, staying under one does not mean being safe, and no label distinguishes the two.
Two different clocks
Acute and chronic toxicity are separate problems with separate time courses, organs and causes, and vitamin A is the one nutrient where ODS lays both out side by side.
Because vitamin A is fat soluble, ODS explains, the body stores the excess, mainly in the liver, where it accumulates. Acute toxicity arrives within days to weeks of one or a few very high doses, typically more than 100 times the recommended allowance, and shows up as severe headache, blurred vision, nausea, dizziness, aching muscles and coordination problems. Chronic hypervitaminosis A, from regular high doses, looks nothing like that: dry skin, painful muscles and joints, fatigue, depression and abnormal liver test results.
The adult limit for preformed vitamin A is 3,000 micrograms of retinol activity equivalents, about 10,000 international units, and its scope has two halves that have to be read together. ODS states that its vitamin A limits apply to both food and supplement intakes, and separately that they cover only products from animal sources and supplements whose vitamin A comes entirely from retinol or its ester forms. So liver at dinner counts, and in a multivitamin that also contains beta-carotene only the retinol portion does. No limit has been set for beta-carotene. Excess preformed vitamin A also causes birth defects, which is why ODS reports experts advise women who are or might be pregnant, and those lactating, against high-dose vitamin A supplements.
Iron is the best-documented acute case, with a regulatory experiment attached. ODS gives the adult limit as 45 milligrams a day and reports that acute intakes above 20 milligrams per kilogram, about 1,365 milligrams for a 150-pound person, can cause corrosive damage to the intestine, with one-time ingestions near 60 milligrams per kilogram leading to organ failure, coma, convulsions and death. Between 1983 and 2000, ODS states, at least 43 American children died after swallowing supplements containing high doses of iron, across a range of 36 to 443 milligrams per kilogram of body weight. FDA began requiring single-dose packaging in 1997 for oral supplements with more than 30 milligrams of elemental iron per dose, and between 1998 and 2002 only one child death from an iron tablet was reported. A court decision led FDA to remove that requirement in 2003, and ODS does not say what happened afterward, so neither can this article.
Poison center data show where acute harm lands. America’s Poison Centers reports that 55 US poison centers handled 2,092,689 human exposures in 2024. Among children aged 5 and under, dietary supplements, herbals and homeopathic products ranked fifth at 59,575 cases, and vitamins seventh at 40,343. Neither category appears among the top ten substances for adults. Both halves matter: acute supplement poisoning is overwhelmingly a child reaching a bottle, while adult excess is slow, cumulative and generates no emergency call. These count exposures, not doses, and no dataset located for this article breaks supplement harm down by how much anyone took.
Excess of one nutrient as deficiency of another
Some excesses do their damage indirectly, by crowding out another nutrient.
Zinc carries a 40 milligram adult ceiling at ODS, which also records that sustaining 50 milligrams or above across several weeks interferes with how the body takes up copper, impairs immune function and depresses HDL cholesterol, while noting that reaching such a figure through food alone is improbable. Such a person does not present with zinc poisoning. They present with the consequences of low copper.
Magnesium makes a different point: the same dose is not the same exposure in different people. ODS gives the supplemental limit as 350 milligrams for adults and is careful about scope, stating these limits include magnesium only from supplements and medications, not magnesium occurring naturally in food and beverages. Article 329 sets out why that limit sits where it does. What matters here is the group ODS singles out: people whose kidneys are not clearing magnesium normally, for whom an identical intake becomes a far larger exposure. Effects have included low blood pressure, vomiting, difficulty breathing, irregular heartbeat and cardiac arrest, and ODS notes a few fatal cases.
An amount that is unremarkable with healthy kidneys can be dangerous without them, and no label accounts for that. It is the strongest argument in this subject for routing a decision to someone who knows a person’s kidney function.
What excess does to the liver, and where dose stops explaining things
The liver data are the most serious numbers available, and only part of that story is about dose.
The Drug-Induced Liver Injury Network enrolled patients prospectively from 2004 through 2013 at eight US referral centers. In the analysis published by Victor Navarro and colleagues in Hepatology in 2014, 130 of 839 enrolled patients, 15.5 percent, were judged to have liver injury caused by herbal and dietary supplements, and the share attributed to supplements rose from 7 percent to 20 percent across the study period. The outcome comparison is what should stop a reader: non-bodybuilding supplement cases led to death or transplantation more often than injury from conventional medications, 13 percent against 3 percent.
The American Association for the Study of Liver Diseases offers an explanation in its 2023 practice guidance that is not about chemistry. Patients with supplement-related liver failure are more likely to die or need a transplant, it says, possibly because of delayed recognition of the product as the cause, or reluctance among supplement consumers to seek care. That points at behavior, which is changeable in a way a toxin is not.
Now the limitation. Most herbal liver injury is idiosyncratic, not dose-dependent. The 2023 guidance links one botanical’s hepatotoxicity to a specific immune-system gene variant, HLA-B*35:01, the signature of an unpredictable individual reaction rather than a dose effect. Folding green tea extract or ashwagandha into a dose-response argument would misdescribe them.
Federal reference material marks the distinction. LiverTox, a reference published by the federal diabetes and digestive diseases institute together with the National Library of Medicine and the injury network itself, grades agents from category A, more than 50 published cases with a characteristic pattern, down to category E, no evidence of liver injury despite extensive use. It also appends a high-dose tag to agents that injure the liver only at high doses, usually as direct toxins. The four examples LiverTox names are aspirin, acetaminophen, niacin and vitamin A. Two of the four are nutrients, which is the dose-dependence of nutrient toxicity written into a federal system.
Dose is also not only the number on the label. A 2020 United States Pharmacopeia expert panel review in Toxicology Reports, with FDA liaisons and liver injury network members among its authors, reported that animal and human data indicate repeated oral bolus doses of green tea extract taken while fasting significantly increase how much reaches the bloodstream, probably by saturating first-pass elimination. The same milligrams, swallowed differently, are not the same exposure.
When the excess was not a choice
One case belongs here because it undercuts the premise that careful label-reading is a sufficient defense.
ODS reports that in 2008, 201 people experienced severe adverse reactions after taking a liquid dietary supplement that contained 200 times the labeled amount of selenium. Nobody in that group took too much on purpose. They took the labeled dose of a product whose label was wrong by more than two orders of magnitude. ODS names neither the product nor its measured content, and neither will this article.
Selenium’s limit is 400 micrograms a day for adults from food and supplements, and ODS notes it does not apply to people taking selenium under a physician’s care. Chronic excess, called selenosis, announces itself through hair loss and brittle or lost nails, with early indicators including a garlic odor on the breath and a metallic taste. Food reaches these amounts too: ODS reports Brazil nuts at 68 to 91 micrograms per nut.
Vitamin D has the same failure mode: ODS reports toxicity cases caused by products containing excessive amounts. Article 347 covers what measurement finds in products.
The signs the agencies describe
None of this is advice about any product, and none of it establishes a dose for anyone. What the sources do supply is a set of recognizable presentations, each tied to a nutrient:
- Tingling, burning or numbness in the hands or feet, with unsteadiness, in anyone taking vitamin B6 at any dose. This is the B6 presentation across all three bodies of evidence, and the one where ODS states recovery is conditional rather than certain, as the earlier section reports.
- Severe headache with blurred vision, days to weeks after unusually large vitamin A intake, which is how the acute syndrome ODS describes begins.
- Nausea, vomiting, muscle weakness or confusion with high-dose vitamin D, the hypercalcemia symptoms ODS lists.
- Unusual bruising or bleeding with high-dose vitamin E, given its effect on platelets and clotting factors.
- Jaundiced skin or eyes, tea-colored urine, pain under the right ribs, or exhaustion with loss of appetite, which is how liver injury declares itself, and where the 2023 guidance blames slow recognition for worse outcomes.
- Hair loss with brittle nails plus a garlic smell on the breath, the selenium picture ODS describes.
- Any of these in a child who may have reached a bottle, where the poison center volume is, and where a poison center call is the documented route.
The takeaway that does the most work is not a number. The solubility heuristic fails in both directions, which means it is not a heuristic at all. Vitamin B6 is water soluble and has disabled people at doses an Australian regulator now considers low enough to restrict. Vitamin K is fat soluble and has no upper limit anywhere, because the Food and Nutrition Board declined to set one, reporting no adverse effects from food or supplement intake in humans or animals. Which side of that split a nutrient sits on predicts nothing about whether more of it is safe.
Sources
- Therapeutic Goods Administration, Australian Government Department of Health. “Notice of final decision to amend the current Poisons Standard: pyridoxine, pyridoxal, pyridoxamine (vitamin B6), 25 November 2025.” tga.gov.au
- Therapeutic Goods Administration, Australian Government Department of Health. “Stronger safety controls to be introduced for products containing vitamin B6.” tga.gov.au
- NIH Office of Dietary Supplements. “Vitamin B6: Fact Sheet for Health Professionals (US upper limit of 100 mg/day; EFSA's 2023 limit of 12 mg/day; sensory neuropathy).” ods.od.nih.gov
- NIH Office of Dietary Supplements. “Niacin: Fact Sheet for Health Professionals (upper limit based on skin flushing; extended-release hepatotoxicity).” ods.od.nih.gov
- NIH Office of Dietary Supplements. “Vitamin A and Carotenoids: Fact Sheet for Health Professionals (acute and chronic hypervitaminosis A; scope of the upper limit).” ods.od.nih.gov
- NIH Office of Dietary Supplements. “Vitamin D: Fact Sheet for Health Professionals (upper limit of 4,000 IU; toxicity unlikely below 10,000 IU).” ods.od.nih.gov
- NIH Office of Dietary Supplements. “Vitamin E: Fact Sheet for Health Professionals (SELECT trial at 400 IU/day; hemorrhagic stroke).” ods.od.nih.gov
- NIH Office of Dietary Supplements. “Iron: Fact Sheet for Health Professionals (acute toxicity in mg/kg; child deaths 1983-2000; packaging history).” ods.od.nih.gov
- NIH Office of Dietary Supplements. “Selenium: Fact Sheet for Health Professionals (2008 mislabeled liquid supplement; selenosis).” ods.od.nih.gov
- NIH Office of Dietary Supplements. “Zinc: Fact Sheet for Health Professionals (copper absorption at 50 mg or more over weeks).” ods.od.nih.gov
- NIH Office of Dietary Supplements. “Magnesium: Fact Sheet for Health Professionals (supplemental upper limit; toxicity risk with impaired kidney function).” ods.od.nih.gov
- NIH Office of Dietary Supplements. “Vitamin K: Fact Sheet for Health Professionals (no Tolerable Upper Intake Level established).” ods.od.nih.gov
- LiverTox, National Institute of Diabetes and Digestive and Kidney Diseases / National Library of Medicine. “Categorization of the Likelihood of Drug Induced Liver Injury (categories A to E and the high-dose designation).” ncbi.nlm.nih.gov
- America's Poison Centers (National Poison Data System). “Poison Statistics: National Data 2024.” poison.org
- Hepatology. “Navarro VJ, Barnhart H, Bonkovsky HL, et al. "Liver injury from herbals and dietary supplements in the U.S. Drug-Induced Liver Injury Network." Hepatology, 2014. PubMed. DOI: 10.1002/hep.27317.” pubmed.ncbi.nlm.nih.gov
- New England Journal of Medicine. “Schaumburg H, Kaplan J, Windebank A, et al. "Sensory neuropathy from pyridoxine abuse. A new megavitamin syndrome." N Engl J Med, 1983. PubMed. DOI: 10.1056/NEJM198308253090801.” pubmed.ncbi.nlm.nih.gov
- Hepatology (American Association for the Study of Liver Diseases). “Fontana RJ, Liou I, Reuben A, et al. "AASLD practice guidance on drug, herbal, and dietary supplement-induced liver injury." Hepatology, 2023. PubMed. DOI: 10.1002/hep.32689.” pubmed.ncbi.nlm.nih.gov
- Toxicology Reports. “Oketch-Rabah HA, Roe AL, Rider CV, et al. "United States Pharmacopeia (USP) comprehensive review of the hepatotoxicity of green tea extracts." Toxicol Rep, 2020. PubMed. DOI: 10.1016/j.toxrep.2020.02.008.” pubmed.ncbi.nlm.nih.gov