Look up magnesium on the National Institutes of Health fact sheet and you will find two numbers that look like they cancel each other out. The fact sheet’s recommended intake for a man of 40 is 420 milligrams a day. The ceiling on the same page is 350.
The ceiling sits below the target. It reads like a typo.
It is not. The two numbers count different things. The recommended amount counts magnesium from everything you swallow. The ceiling counts only what arrives in a pill or a medicine. Once you see that, a lot of confusing nutrition advice stops being confusing.
That distinction runs through the whole system of upper limits, and it is not applied the same way twice. For most nutrients the ceiling covers food, water and supplements together. For a few it covers only the manufactured version. For one of them it covers a single chemical form across every source at once. Telling those apart is the entire skill, and it is the reason this article exists.
What an upper limit is
The formal name is the Tolerable Upper Intake Level, usually shortened to UL. It belongs to the Dietary Reference Intakes, the family of population figures developed by the Food and Nutrition Board at the National Academies of Sciences, Engineering, and Medicine. The NIH Office of Dietary Supplements republishes them on its fact sheets, which is where most readers meet them. Article 322 handles the rest of the family, including the Estimated Average Requirement and the Recommended Dietary Allowance.
A UL is the highest daily intake the board judged likely to pose no risk of adverse health effects to almost everyone in the general population. That is a more cautious claim than it sounds, and a narrower one.
The 1998 Institute of Medicine report that set out how these figures get built makes four points that rarely survive the trip into consumer writing. A UL is not in itself a description of risk to any individual. Setting one does not mean that intakes above the recommended amount are being recommended as useful. Risk above the ceiling rises gradually as intake rises, rather than switching on at the line. And the UL is not meant to apply to people who are being treated with the nutrient under medical supervision.
Put those together and you get something more useful than a maximum safe dose. A UL marks the top of the zone where a committee judged the risk of harm to be indistinguishable from zero, for healthy people, taking it every day, without a clinician involved. Change any one of those conditions and the number is answering a different question than the one you are asking.
The figures matter because the exposure is ordinary. Most American adults, and roughly a third of children and adolescents, had taken at least one supplement in the previous month when CDC last counted. Article 321 carries the survey figures.
The default rule is food, water and supplements together
The Food and Nutrition Board’s own upper-limit tables carry a short note that settles most cases before you start: “Unless otherwise specified, the UL represents total intake from food, water, and supplements.”
That default governs the large majority of nutrients, and the Office of Dietary Supplements spells it out on individual fact sheets. The zinc ULs cover zinc from food and supplements, and the adult figure for ages 19 and older is 40 mg a day. The iron ULs cover food and supplements, 45 mg a day for adults, and are based on the gastrointestinal effects seen after supplemental doses of iron salts. The vitamin B6 ULs apply to both food and supplement intakes, 100 mg a day for adults. Vitamin C and selenium work the same way.
For all of those, a fortified breakfast cereal, a glass of fortified juice and a capsule go into the same running total. There is no separate supplement allowance sitting on top of what you eat.
That is the rule. Now the exceptions, which are where nearly every mistake in this subject lives.
The five exceptions, and the one that breaks the pattern
Five nutrients carry a UL that does not cover everything you consume. Four of them are exceptions in the same direction. The fifth is not, and it is the one most likely to be written up wrong.
Magnesium. The Office of Dietary Supplements states that the magnesium ULs cover only supplements and medications, not magnesium occurring naturally in food and drink. The figure is 350 mg a day for everyone aged nine and up, including pregnancy and lactation.
Folate. The board set a ceiling for the synthetic forms of folate found in dietary supplements and fortified foods, and set none at all for folate from food. The adult figure is 1,000 mcg a day. Watch the units here, because the fact sheet switches them mid-page: the ULs are in micrograms of the synthetic form, while the recommended intakes are in micrograms of dietary folate equivalents, a different scale.
Vitamin e. The ULs apply to all forms of supplemental alpha-tocopherol, including the eight stereoisomers present in synthetic vitamin E. The adult figure is 1,000 mg a day. ODS makes a second, separate statement on the same page about large intakes from food, and article 327 reports it alongside the rest of the vitamin E evidence.
Niacin. ODS says the ULs apply only to supplemental niacin, covering both nicotinic acid and nicotinamide, at 35 mg a day for adults. The recommended intake for an adult man is 16 mg of niacin equivalents, which is again a different unit from the ceiling.
The vitamin A trap
Vitamin A is where the pattern breaks, and anyone who has learned that form-specific means supplements only will get it backwards.
The Office of Dietary Supplements states that the Food and Nutrition Board set ULs for preformed vitamin A that apply to both food and supplement intakes. Preformed vitamin A means retinol and its ester forms, such as retinyl palmitate, which is what you get from animal foods and from most vitamin A in supplements. The board set no ULs for beta-carotene or the other provitamin A carotenoids, the orange pigments in plants that your body converts as needed.
So the vitamin A ceiling is specific about form and indifferent about source. The adult figure is 3,000 mcg a day of preformed vitamin A. A serving of liver counts toward it. A carrot does not. Every other exception on this list restricts the source; this one restricts the chemistry and then counts everything.
The stakes are not abstract. ODS reports that intakes of preformed vitamin A above the UL, along with some retinoid medications, can cause congenital birth defects including malformations of the eye, skull, lungs and heart, and that experts advise women who are or might become pregnant to avoid high-dose supplements of more than 3,000 mcg RAE, or 10,000 IU, a day. That is a published position of expert bodies as ODS reports it, directed at a defined group, and it is the kind of question an obstetric clinician answers for an individual.
Two federal sources, two answers
Two of those five entries are described differently by two authoritative federal sources, and neither disagreement has been resolved.
On niacin, the Office of Dietary Supplements says the ULs apply only to supplemental niacin, and the page does not mention fortified foods at all. The Food and Nutrition Board’s own table footnote says the ULs for vitamin E, niacin and folate apply to synthetic forms obtained from supplements, from fortified foods, or from a combination of the two. A niacin-fortified breakfast cereal counts toward the ceiling under the board’s wording and does not count under the ODS wording.
On magnesium, ODS says supplements and medications. The board’s element footnote says the magnesium ULs represent intake from a pharmacological agent only and do not include intake from food and water. A magnesium capsule bought off a shelf is plainly a dietary supplement. Whether it is a pharmacological agent is a separate question, and the two sources give you two different categories to sort it into.
Both sources are federal, both are primary, and both are current. This article reports the disagreement rather than picking a side, because picking a side would mean inventing a resolution that neither body has published.
Why the magnesium ceiling sits below the magnesium target
Magnesium is the clearest demonstration available of what a source-specific ceiling does to the arithmetic, and ODS explains it on its own page rather than leaving readers to work it out.
The fact sheet states: “The ULs only include magnesium from dietary supplements and medications; they do not include magnesium found naturally in food and beverages.”
It then notes directly that for many age groups the UL appears to be lower than the recommended intake, and explains that this happens because the recommended intakes count magnesium from all sources, food, beverages, supplements and medications together.
The numbers make it concrete. ODS lists recommended intakes of 400 mg for men aged 19 to 30 and 310 mg for women that age, rising to 420 mg for men and 320 mg for women from 31 onward. The ceiling for everyone aged nine and up is 350 mg. A 40-year-old woman whose recommended intake is 320 mg and whose supplement ceiling is 350 mg is not being told to eat 320 and stop at 350. She is being given two numbers measured from two different places.
What sits behind the 350 figure is also worth knowing, because it is not organ damage. ODS attributes diarrhea, nausea and abdominal cramping to high doses from supplements or medications. Genuine magnesium toxicity, with low blood pressure, vomiting, difficulty breathing, irregular heartbeat and cardiac arrest, is linked to very large amounts, typically those providing more than 5,000 mg a day, usually from magnesium-containing laxatives and antacids. Kidney function moves the whole question, and ODS says so on the same page; article 329 works through the magnesium case in detail. For anyone whose kidneys are not clearing normally, the published ceiling is not the operative number, and what replaces it belongs with the clinician managing that condition.
A missing upper limit means two different things
Several nutrients have no UL at any life stage. On the Food and Nutrition Board’s tables that includes vitamin K, thiamin, riboflavin, vitamin B12, pantothenic acid, biotin and the carotenoids among the vitamins, and arsenic, chromium, silicon and sulfate among the elements, along with potassium and sodium.
Here is the part worth slowing down for. The same table uses the abbreviation ND, for not determinable, with two different footnotes attached.
The first footnote covers most nutrients: not determinable because of a lack of data on adverse effects in that age group, and concern about the ability to handle excess amounts, with the instruction that intake should come from food only. That is an absence of evidence combined with an active worry.
The second footnote, used for potassium and sodium, says not determinable owing to a lack of a specific toxicological effect. That is an absence of something to measure.
Those two point in opposite directions and the table does not flag the difference. ODS’s own wording shows the same split. On vitamin B12 it says the board did not set a UL because of the vitamin’s low potential for toxicity. On chromium it says the board concluded that no adverse effects have been linked to high intakes from food or supplements and so set no UL, but then adds that caution may be warranted because the data are limited and high intakes could have adverse effects. Those are two different evidentiary situations wearing the same label.
The 1998 Institute of Medicine methods report settles the general principle. Where no ceiling could be derived, adverse effects should still be assumed possible at some level of intake. A missing ceiling is a statement about what the committee could measure. It is not a safety rating, and it is certainly not a finding that any amount is fine.
One more wrinkle: for infants under 12 months, many nutrients carry the ND marking with the same instruction, that the source of intake should be food only. That is the most restrictive reading in the tables, and it applies to the group least able to tell anyone that something is wrong.
When crossing the ceiling is the plan
The clearest case of a popular regimen built to exceed a published ceiling is the zinc lozenge.
The zinc UL for adults is 40 mg a day and it counts food and supplements together. On the same ODS fact sheet, the section on the common cold reports an earlier review finding that lozenge doses above 75 mg a day shortened colds while lower doses did not, and a 2021 meta-analysis using total daily doses of 45 to 300 mg. The efficacy floor in that literature is 1.875 times the ceiling. The top of the trial range is 7.5 times it.
That is not a scandal and it is not a loophole. It is a short course of a high dose used for a few days for a specific purpose, which is a different exposure from the daily intake sustained indefinitely that the ceiling figure describes. The evidence on whether zinc lozenges work is genuinely contested and article 331 handles it properly. What belongs here is the arithmetic, because a reader who believes a UL is a hard safety line will not be able to make sense of a protocol that doubles it on purpose.
Prescribing does the same thing from the other direction. ODS states plainly that physicians sometimes prescribe iron intakes higher than the UL, such as when people with iron-deficiency anemia need higher doses to replenish their stores. The carve-out for medical supervision appears on fact sheet after fact sheet: folate, niacin, zinc, selenium, vitamin B6 and iodine all carry a version of it.
So the ULs are a description of unsupervised, chronic, daily intake by healthy people. A short course, a supervised course, or a person with a condition that changes how the nutrient is handled all fall outside what the number was built to describe. Deciding which of those you are in is not something a label can do for you, and it is the specific question to put to a clinician or pharmacist rather than a general one about whether supplements are safe.
Reading a ceiling figure without getting it wrong
The population data show why this is not a theoretical exercise. CDC’s 2026 National Nutrition Report found that vitamin D insufficiency in people aged six and older fell from 28 percent to 22 percent between 2001 to 2002 and August 2021 to August 2023, while the share with levels above 125 nmol/L rose from under 1 percent to 8 percent, increasing most in adult supplement users, women and non-Hispanic White people. Movement in both directions at once.
A 2012 analysis of national survey data by Bailey and colleagues at the Office of Dietary Supplements, published in the Journal of the Academy of Nutrition and Dietetics, found the same trade-off inside one dataset: supplement use lowered the share of adults falling below the Estimated Average Requirement for every vitamin examined, and raised the likelihood of intakes above the UL for folic acid and vitamins A, B6 and C. Both halves of that sentence are true at once.
A 2016 CDC-authored analysis in Nutrients, by Orozco and colleagues, puts a number on the folate rule. Among 18,321 non-pregnant US adults in surveys from 2003 to 2010, 2.7 percent had usual daily folic acid intakes above the ceiling, 98.5 percent of them took supplements containing folic acid, and nobody in the study population exceeded it from mandatorily fortified enriched grain products alone. That is the scope rule showing up in real data.
Harms are harder to pin down. A 2019 cohort study in Annals of Internal Medicine by Chen and colleagues linked calcium intake above the UL to higher cancer mortality, and the association appeared to track supplemental calcium rather than calcium from food. That is observational evidence, the authors name residual confounding as a limitation, and it is not proof of cause.
When you meet a ceiling figure anywhere, five questions get you to the right reading of it:
- Who published it, and did you read it there? ODS and the Food and Nutrition Board word two of these differently.
- Which age and sex band is it for? Infant, child, adolescent and pregnancy figures differ, sometimes by a factor of five.
- What does it count? Everything, supplements only, or supplements plus fortified food.
- Which form and which unit? Micrograms are not micrograms DFE, milligrams are not milligrams NE, and micrograms are not IU.
- Does the supervision carve-out apply? The figure was not built for people treated with the nutrient by a clinician.
The single most useful habit is refusing to compare any number to a UL until you know what the UL counts. A label line reading 100% Daily Value tells you nothing about where you sit against one of these figures, because the Daily Value and the UL are unrelated numbers built for different jobs. Article 324 explains what the label number is.
Sources
- National Academies of Sciences, Engineering, and Medicine, Food and Nutrition Board. “Dietary Reference Intakes: Tolerable Upper Intake Level tables for vitamins and elements, with table notes and footnotes.” nationalacademies.org
- Institute of Medicine, National Academies Press (via NCBI Bookshelf). “A Model for the Development of Tolerable Upper Intake Levels, Chapter 3 of Dietary Reference Intakes for Thiamin, Riboflavin, Niacin, Vitamin B6, Folate, Vitamin B12, Pantothenic Acid, Biotin, and Choline (1998).” ncbi.nlm.nih.gov
- NIH Office of Dietary Supplements. “Magnesium: Fact Sheet for Health Professionals.” ods.od.nih.gov
- NIH Office of Dietary Supplements. “Niacin: Fact Sheet for Health Professionals.” ods.od.nih.gov
- NIH Office of Dietary Supplements. “Vitamin A and Carotenoids: Fact Sheet for Health Professionals.” ods.od.nih.gov
- NIH Office of Dietary Supplements. “Folate: Fact Sheet for Health Professionals.” ods.od.nih.gov
- NIH Office of Dietary Supplements. “Vitamin E: Fact Sheet for Health Professionals.” ods.od.nih.gov
- NIH Office of Dietary Supplements. “Zinc: Fact Sheet for Health Professionals.” ods.od.nih.gov
- NIH Office of Dietary Supplements. “Iron: Fact Sheet for Health Professionals.” ods.od.nih.gov
- NIH Office of Dietary Supplements. “Vitamin B12: Fact Sheet for Health Professionals.” ods.od.nih.gov
- NIH Office of Dietary Supplements. “Chromium: Fact Sheet for Health Professionals.” ods.od.nih.gov
- Centers for Disease Control and Prevention. “2026 National Nutrition Report: At A Glance (released 24 June 2026).” cdc.gov
- PubMed. “Bailey RL, Fulgoni VL, Keast DR, Dwyer JT. "Examination of vitamin intakes among US adults by dietary supplement use." Journal of the Academy of Nutrition and Dietetics, 2012. DOI: 10.1016/j.jand.2012.01.026.” doi.org
- PubMed. “Orozco AM, Yeung LF, Guo J, Carriquiry A, Berry RJ. "Characteristics of U.S. Adults with Usual Daily Folic Acid Intake above the Tolerable Upper Intake Level: National Health and Nutrition Examination Survey, 2003-2010." Nutrients, 2016. DOI: 10.3390/nu8040195.” doi.org
- PubMed. “Chen F, Du M, Blumberg JB, et al. "Association Among Dietary Supplement Use, Nutrient Intake, and Mortality Among U.S. Adults: A Cohort Study." Annals of Internal Medicine, 2019. DOI: 10.7326/M18-2478.” doi.org