Too much vitamin A: where the limit really is
Futures Nutrition Editorial Team · 14 August 2026
Too much vitamin A: where the limit really is
14 August 2026
The short answer: the tolerable upper intake level for preformed vitamin A is 3,000 µg a day for adults — that has applied since 2002 and was confirmed by EFSA in June 2024 after a full re-evaluation. The figure covers total intake from everything: food, fortified products, tablets. In practice hardly anyone exceeds it through supplements, but through a single food: liver. In German food monitoring, 100 grams of slaughter-animal liver contained an average of 17 to 29 mg of vitamin A — six to ten times the daily upper limit in half a portion. A tablet with 400 µg, by contrast, comes to about 13 per cent of it. If you want to keep one rule in mind, make it this one: liver at most once a month, and not at all in pregnancy.
With hardly any other vitamin do the perceived and the actual risk diverge so far. Vitamin A is fat-soluble, is stored in the liver and can accumulate — that much is true. It does not follow that the capsule is the problem.
The upper limit: one figure, nine age groups
The authoritative document is the opinion of the EFSA Panel on Nutrition, Novel Foods and Food Allergens of 6 June 2024 (EFSA Journal 2024;22(6):e8814). It systematically re-examines the values set by the former Scientific Committee on Food (SCF) in 2002 — for harm to the unborn child, liver toxicity and bone health — and retains them.
| Age group | Upper limit (UL) per day |
|---|---|
| Infants 4–11 months | 600 µg RE |
| Children 1–3 years | 800 µg RE |
| Children 4–6 years | 1,100 µg RE |
| Children 7–10 years | 1,500 µg RE |
| Adolescents 11–14 years | 2,000 µg RE |
| Adolescents 15–17 years | 2,600 µg RE |
| Adults from 18 years | 3,000 µg RE |
| Pregnant women | 3,000 µg RE |
| Breastfeeding women | 3,000 µg RE |
Three points sit in the small print and matter more than the figures themselves.
The value applies to all sources together. It covers foods, fortified products and food supplements — and all forms authorised for foods: retinol, retinyl acetate, retinyl palmitate.
It applies only to preformed vitamin A. β-carotene does not fall under it; more on that below.
The values for children are calculated, not measured. They are derived from the adult value by allometric scaling using body weight to the power of 0.75. That is standard practice, but it means there are no separate data for children, only a scaling.
Why the limit hangs on pregnancy
Out of three possible adverse effects, EFSA selected one as decisive: harm to the unborn child. Not because it is the most frequent, but because it is irreversible and occurs at a comparatively low intake — a limit that protects against it also covers the other effects.
The figure comes from a prospective cohort study in the United States. Rothman and colleagues surveyed 22,755 pregnant women in weeks 15 to 20 about their retinol intake from food and supplements during the twelve weeks after their last menstrual period (N Engl J Med 1995;333(21):1369–1373, PMID 7477116). Among women with at least 4,501 µg RE a day from all sources, the proportion of newborns with malformations was raised (prevalence ratio 2.2; 95% CI 1.3–3.8), and more clearly so for cranial neural-crest defects (3.5; 1.7–7.3). For intake from supplements alone the ratio was 2.4 and 4.8 respectively. In the lower intake categories the proportion remained stable — the dose-response analysis indicated a threshold at 3,000 µg RE a day.
Those 3,000 µg have been the upper limit ever since, without an additional safety factor: EFSA considers it plausible that the true threshold lies higher, because in a European cohort hardly any cases occurred even at median doses of 15,000 µg RE from supplements. Five case-control studies support the assessment.
In practice this means: the critical window is the first trimester and the weeks before it — the phase in which many women do not yet know they are pregnant. The German Nutrition Society therefore advises pregnant women, especially in the first trimester, to avoid regular and excessive consumption of foods very rich in vitamin A, liver above all. EFSA puts it more narrowly still: women who are pregnant or intend to become pregnant should not eat liver, offal or products made from them at all.
Liver: the one source that breaks the limit
The figure on which everything turns comes from the 2016 food monitoring programme of the German Federal Office of Consumer Protection and Food Safety: liver from slaughter animals contains an average of between 17 and 29 mg of vitamin A per 100 g. The Federal Institute for Risk Assessment (BfR) cites this range in its maximum-level assessment.
| Food or supplement | Vitamin A per portion | Share of the upper limit (3,000 µg) |
|---|---|---|
| Calf's liver, 125 g | approx. 21,000–36,000 µg | 700–1,200 % |
| Liver sausage, 30 g on bread | approx. 1,000–2,500 µg | 35–85 % |
| Tablet with 400 µg vitamin A | 400 µg | 13 % |
| Butter, 20 g | approx. 130 µg | 4 % |
| Carrots, 200 g (as β-carotene) | not covered by the UL | — |
Liver values from the 2016 food monitoring programme; liver sausage and butter from USDA FoodData Central (4,090–8,310 µg and 671 µg retinol per 100 g respectively). The ranges are real: the vitamin A content of liver varies with species, age and feeding.
A single portion of liver therefore delivers seven to twelve times what counts as the tolerable daily intake. That is no reason to panic — the UL refers to sustained daily intake, not to a single plate. EFSA spells this out: European populations do not exceed the value on average as long as "liver, offal and products thereof" appear on the table at most once a month. Anyone who eats offal more often, or additionally takes high-dose supplements, may end up above it.
What actual intake in Germany looks like is shown by the second National Nutrition Survey. Looking at preformed vitamin A alone, adults took in a median of 0.4 to 0.5 mg (women) and 0.6 to 0.7 mg (men) a day. At the 95th percentile — the high consumers — it was 1.3 to 2.0 mg for women and 2.5 to 2.9 mg for men. The median sits comfortably below the limit; the upper end of the distribution touches it.
What the BfR concludes from this for supplements
From that distribution the BfR draws an uncomfortable conclusion: between the upper limit and what high consumers eat anyway, nothing is left over on paper. For adolescents aged 15 to 17 the calculation UL minus 95th percentile comes to exactly zero. The institute therefore names two options for the use of preformed vitamin A in food supplements: either no addition — or an addition of up to 0.4 mg a day, which, allowing for an uncertainty factor of 2 for multiple use, corresponds to a maximum level of 0.2 mg per recommended daily portion. It additionally recommends that vitamin-A-containing supplements carry a note that vitamin A should be taken in pregnancy only after consulting a doctor.
That is a proposal for rule-making, not a legally binding maximum — binding EU maximum levels for vitamins in food supplements still do not exist. Anyone comparing supplements should nevertheless know the figure. Our Vitamin A + E + lutein contains 400 µg of vitamin A per tablet (50 % of the 800 µg reference value) and thus lies above the BfR proposal, but at around one eighth of the upper limit. What that amount means in terms of the approved claim is set out in Vitamin A and vision: the approved claim.
The flip side of the same assessment is interesting: for the fortification of ordinary foods the BfR considers preformed vitamin A unsuitable — with one exception. Margarine and mixed fat products sold as a butter substitute have had to be fortified with vitamin A in Germany for decades; here a recommended maximum level of 1 mg per 100 g applies.
µg, IU and RE: why the same amount looks different three times
Part of the confusion arises on the label. Older packs and many imported products declare vitamin A in International Units (IU), and the conversion factor depends on the chemical form.
| Form | 1 mg corresponds to | in retinol equivalents |
|---|---|---|
| Retinol | 3,330 IU | 1,000 µg RE |
| Retinyl acetate | 2,900 IU | 870 µg RE |
| Retinyl palmitate | 1,830 IU | 550 µg RE |
In practice that means: 10,000 IU from retinyl palmitate is around 3,000 µg RE — precisely the upper limit, in a single capsule. Such products exist on the market, mostly from countries with different regulation. Anyone who sees an IU figure should convert before comparing.
Then there is the second key, which has nothing to do with IU: retinol activity equivalents (RAE) convert 12 µg of β-carotene into 1 µg of retinol, while the older retinol equivalents (RE) use only 6 µg. For pure retinol the two figures are identical; for plant sources this produces the factor of two between two tables that mean the same vegetable.
β-carotene: no upper limit, but a clear restriction
For β-carotene, EFSA in 2024 was able to establish neither an upper limit nor a safe level of intake — and that is expressly not a free pass but a data gap. The panel selected lung cancer risk as the critical effect; robust data exist only for 20 mg a day from supplements, and below that dose there is not a single suitable study. Without a dose-response curve no reference point can be determined.
What the authority says instead is concrete enough:
- From a normal diet, β-carotene raises no concern — including as a food colour. Fruit and vegetables are the main sources, and there is no indication of an adverse effect from them.
- Smokers should avoid supplements containing β-carotene. In the ATBC study in heavy smokers, lung cancer risk rose statistically significantly on 20 mg a day over six years.
- For everyone else: only to cover the vitamin A requirement. With the 6:1 factor, the EFSA intake recommendations correspond to about 4.5 mg of β-carotene for men and 3.9 mg for women.
- β-carotene does not harm the unborn child. The body throttles absorption and conversion once the stores are full.
For food supplements the BfR proposes a maximum level of 3.5 mg of β-carotene per recommended daily portion — an order of magnitude that fits those figures.
How an oversupply shows itself
Liver toxicity is the most serious but late consequence of a chronically excessive intake. The lowest dose at which it has been described in humans is 7,500 µg RE a day over six years — two and a half times the value on which the upper limit is based. The range extends from reversibly raised liver enzymes to fibrosis and cirrhosis. EFSA notes that liver enzymes are an unreliable early indicator here: local fibrosis can precede the rise.
Two further observations concern considerably higher amounts: a bulging fontanelle in infants and unfavourable changes in blood lipids in adults, both from about 7,500 µg RE as a single or repeated dose. And a formulation question that is rarely mentioned: water-miscible retinyl esters are absorbed faster than oil-based ones and are considered about ten times more toxic at the same amount.
What is missing is a simple laboratory value. Serum retinol is held constant across wide ranges and does not reliably indicate an oversupply; EFSA expressly names the development of non-invasive markers as an open research task.
Food supplements are not a substitute for a balanced and varied diet and a healthy lifestyle.
The 1.5 mg recommendation for post-menopausal women
Anyone encountering older sources will find a second, stricter figure there: 1,500 µg a day for women after the menopause. It comes from the 2002 SCF assessment and was based on indications that a high vitamin A intake might be connected with bone fractures. Even then the committee expressly called for better data.
EFSA evaluated those data in 2024 — and concludes that a causal connection between an intake of 1,000 to 3,000 µg RE a day and an increased fracture risk cannot be substantiated with the available evidence. The evidence added since 2002 does not support the link. Consistently, the upper limit of 3,000 µg expressly applies to post-menopausal women as well.
That is practically relevant because German information sheets still carry the old figure — the 2021 BfR paper, for instance, names it and points out that post-menopausal women at the 95th percentile are already above it, at 1.7 to 2.0 mg, from ordinary food alone. Under the current EU assessment that exceedance is no longer a finding. Anyone reading the two figures side by side is not reading an oversight but two states of assessment.
What this means at the shelf
- Add up the day, not the tablet. The UL applies to everything together. Vitamin A is also in multi-ingredient supplements and in butter or margarine.
- Convert IU. 10,000 IU of retinyl palmitate is around 3,000 µg RE — the full upper limit in one capsule.
- Treat liver as an event, not a habit. Once a month is unremarkable; in pregnancy and when trying to conceive, not at all.
- With β-carotene, mind the smoking. Smokers leave β-carotene supplements aside — including those marketed as "natural".
- Know the second carrier of the vision claim. Zinc carries it as well and has a quite different safety margin, for instance in Zinc + vitamin B3. Which nutrients may carry the claim at all is set out in Eye nutrients: what is approved and what is not; what else the range holds on the subject is shown by the category Eyes.
Frequently asked questions
Is a tablet with 400 µg of vitamin A a day too much? Measured against the upper limit of 3,000 µg a day, no — that is around 13 per cent. Measured against the BfR proposal of 0.2 mg per recommended daily portion, it lies above it. The two figures answer different questions: the UL says from what point an adverse effect is to be expected; the BfR value says how much a single supplement should contribute so that high consumers and multiple users also stay in the safe range. Anyone taking further vitamin-A-containing supplements adds them up.
Can you poison yourself with carrots? No. β-carotene does not fall under the upper limit for preformed vitamin A, and conversion is throttled once enough vitamin A is present. At very high intakes over months the skin can turn a yellowish-orange (carotenodermia) — that is harmless and reverses. The EFSA restriction applies to supplements, not to the vegetable. What role the related carotenoids lutein and zeaxanthin play is set out in Lutein and zeaxanthin: origin, occurrence, evidence.
I was pregnant and ate liver sausage once — is that a problem? A single portion is something other than a habit. The upper limit describes sustained daily intake, and the malformation data come from an analysis of regular intake over weeks. The recommendation to avoid liver in pregnancy nevertheless stands unchanged — it is worded so strictly because the vitamin A content of liver varies widely and nobody can estimate at the table how much is in it. If in doubt, medical care is the right place to turn, not your own arithmetic.
Why does β-carotene have no upper limit even though there is a risk? Because an upper limit presupposes a dose-response curve, and that is missing. There are data for 20 mg a day in smokers and not a single suitable study below that. EFSA was therefore able to name neither a UL nor a "safe" amount — and instead formulated two rules of conduct: smokers avoid supplements, everyone else limits intake to what covers the vitamin A requirement.
