Red Dye 3 Banned: Food Colour Rules Worldwide (2026 Update)

In January 2027, a synthetic dye that has been illegal in American cosmetics since 1990 will finally become illegal in American food.

That sentence contains the entire strangeness of how colour additives have been regulated. FD&C Red No. 3 — erythrosine — was removed from lipstick thirty-five years before it was removed from sweets, under the same law, on the same evidence, by the same agency. The compliance deadline for food is January 2027 and for ingested drugs January 2028.

Red 3 is the most advanced case, but it is not isolated. The FDA has stated an intention to remove petroleum-based dyes from the American food supply and to approve replacements from natural sources. From 2026 manufacturers may use a "No Artificial Colors" claim provided they are not using petroleum-based colours. Meanwhile the European Union has required warning labels on several dyes since 2010, and many manufacturers reformulated rather than carry them.

This guide covers which dyes are affected where, what the evidence actually shows, what the 2027 deadline means in practice, and how to identify synthetic colours on a label in any country.

The regulatory position varies by dye and by jurisdiction. This table reflects status as of 2026.

Two rows deserve particular attention because they run in the opposite direction to the usual narrative.

Quinoline Yellow and Ponceau 4R are permitted in the EU but not approved for food use in the United States. The divergence is not uniformly one-directional, and articles claiming Europe is categorically stricter on colours are overstating a real but uneven pattern.

Fast Green and titanium dioxide run the other way — permitted in America, not in Europe. Titanium dioxide is technically a colour additive rather than a dye, providing whiteness and opacity rather than hue, and it is the largest single outstanding divergence in this category.

This is the case that best illustrates how regulatory inertia operates in practice.

Erythrosine is a synthetic cherry-red dye, used in sweets, baked goods, maraschino cherries and similar products. In the 1980s, studies found thyroid tumours in male rats exposed to high doses. That finding engaged the Delaney Clause, a provision of American law stating that no food additive shown to induce cancer in humans or animals may be deemed safe.

In 1990 the FDA acted on that finding — and banned Red 3 from cosmetics and externally applied drugs only. The food authorisation remained in place. The result was a situation that was legally coherent only in a narrow procedural sense and indefensible in any practical one: a substance too risky to apply to skin remained permitted to eat.

The explanation lies in process rather than science. Revoking a colour additive listing requires a formal rulemaking process, which requires agency resources, survives industry comment, and is vulnerable to legal challenge. Cosmetic and food authorisations sat under separate listings, and acting on one did not compel action on the other. Absent sustained external pressure, the food listing simply persisted.

That pressure eventually arrived through petitions and public campaigning, and in January 2025 the FDA revoked the food authorisation. Manufacturers have until January 2027 for food and January 2028 for ingested drugs.

The generalisable lesson is not about erythrosine specifically. It is that in a system where removal requires affirmative action and continuation requires none, the default outcome is continuation — even where the agency has already accepted the underlying evidence in another context.

The most common claim about synthetic dyes concerns children's behaviour, and it deserves careful handling because it is neither baseless nor as settled as often presented.

The pivotal research was conducted at the University of Southampton and published in 2007. It examined mixtures of synthetic colours combined with the preservative sodium benzoate in children from the general population, and reported increased hyperactive behaviour associated with the mixtures.

Regulators drew different conclusions from the same study, which is instructive.

The European Union responded by requiring a warning on foods containing any of six specific dyes: Tartrazine (E102), Quinoline Yellow (E104), Sunset Yellow (E110), Carmoisine (E122), Ponceau 4R (E124) and Allura Red (E129). The warning states that the colour "may have an adverse effect on activity and attention in children". Notably the EU did not ban these dyes — it required disclosure and left the choice to manufacturers and consumers.

That regulatory choice produced a market effect larger than the rule itself. Rather than carry a warning label, many manufacturers reformulated with natural colours. The warning functioned as a de facto phase-out, which is a useful demonstration that mandatory disclosure can achieve what prohibition would have achieved, with less regulatory conflict.

The FDA's assessment at the time was that the evidence did not establish a causal relationship for the general population, while acknowledging that some children may be sensitive. Subsequent reviews, including work by California state scientists, have generally concluded that synthetic dyes can affect behaviour in some children, while noting substantial variation in individual susceptibility and limitations across the literature.

The honest position is that the effect appears real for a subset of children, is not uniform across the population, and is modest in the aggregate — while also being entirely avoidable, since these dyes serve no function other than appearance.

Synthetic colours cluster in predictable categories, and the pattern says a good deal about who they are aimed at.

Two observations follow from this list.

Child-directed products are heavily represented, which is the crux of the regulatory debate. The behavioural evidence concerns children specifically, and the products most likely to contain multiple synthetic dyes are exactly the products marketed to children. That overlap is not coincidental — bright colour is an effective marketing tool for that audience, which is precisely why it is used.

And the last row matters more than people expect. Ingested drugs and supplements carry colour additives for tablet coating and dose identification, which is why the Red 3 revocation includes a separate and later deadline of January 2028 for that category. Medication reformulation involves regulatory processes that food reformulation does not.

A third pattern worth noting is restoration. Several uses on this list exist to replace colour destroyed during processing — heat, light and time degrade natural pigments. The dye is compensating for what industrial processing removed, which is a reasonable summary of a great many additives.

Between now and January 2027 a large-scale reformulation is happening across the American food supply, and it is visible if you know what to look for.

Products containing Red 3 must be reformulated, reconstituted with a permitted alternative, or withdrawn. For most products the replacement is straightforward: Red 40 is a permitted synthetic alternative, and natural options such as beetroot-derived colours, anthocyanins from fruit and vegetables, carmine from cochineal insects, and paprika extract are available.

Each replacement carries trade-offs that explain why reformulation is not instant. Natural colours are generally less stable under heat, light and pH variation than synthetics, which affects shelf life and appearance consistency. Some carry their own flavour. Beetroot-derived reds shift toward brown under heat. Carmine is not suitable for vegetarian, vegan, kosher or halal products. Cost is typically higher.

The FDA's 2026 agenda addresses the supply side of this directly, including publishing draft guidance on when fruit- and vegetable-derived juices qualify as colour additives — a technical question that determines how easily manufacturers can use these alternatives without a separate additive approval.

For a shopper the practical implication is that American products are changing quietly. A product you buy in late 2026 may have a different colour formulation than the same product bought a year earlier, and in most cases nothing on the packet will announce it. If you see Red 3 on an American label today, you are looking at a formulation with a legal expiry date.

The "No Artificial Colors" claim permitted from 2026 is the one change you will see announced, because it is a marketing advantage rather than a compliance obligation.

"Natural colour" is not automatically a better answer, and understanding the alternatives explains why reformulation takes time and why some products resist it.

The pattern running through the limitations column is the whole story. Synthetic dyes exist because they solve these problems: they are stable across heat, light, pH and time, they are flavourless, they are consistent between batches, and they are inexpensive. Every natural alternative sacrifices at least one of those properties.

This has consequences beyond manufacturer convenience. A product coloured with anthocyanins may look different at the end of its shelf life than at the beginning. One coloured with beetroot cannot be baked at high temperature without shifting brown. Spirulina blue is genuinely difficult to work with, which is why blue remained a synthetic-only colour for far longer than red or yellow.

Two further points deserve flagging because they complicate the simple natural-is-better framing.

Carmine is entirely natural and derived from insects, which makes it unsuitable for vegetarian, vegan, kosher and halal diets and a documented if uncommon allergen. Natural origin and suitability are separate questions.

And "natural" is not a strictly defined term for colours in most jurisdictions. Some plant-derived colours are heavily processed, and the FDA's 2026 commitment to publish draft guidance on when fruit- and vegetable-derived juices qualify as colour additives exists precisely because that boundary is currently unclear.

Several jurisdictions are moving simultaneously, with different mechanisms.

The United States is the most active. Beyond the Red 3 revocation, the FDA has stated an intention to remove petroleum-based dyes from the food supply and to approve new colour additives from natural sources. Its 2026 agenda includes draft guidance on fruit- and vegetable-derived juices as colour additives. State legislatures have moved independently — California has passed legislation covering several additives including Red 3 ahead of federal action, and other states have introduced similar bills.

That state-level activity has a disproportionate effect. Because maintaining separate formulations for individual states is commercially impractical, a large state acting alone tends to set a de facto national standard. Manufacturers reformulate nationally rather than regionally.

The European Union continues with its warning-label approach for the six Southampton dyes, which has already driven substantial voluntary reformulation. Its broader additive re-evaluation programme continues to work through substances authorised under older standards.

The United Kingdom retains EU-derived colour rules following its departure from the bloc, with the Food Standards Agency responsible for any divergence.

The underlying commercial logic pushes in one direction regardless of regulation. Synthetic dyes are cheap, stable and effective, but "no artificial colours" has become a marketing asset. Where consumers will pay for the claim, manufacturers reformulate without being compelled — which is why the EU warning label proved so effective, and why the American "No Artificial Colors" claim permitted from 2026 may accelerate change faster than the Red 3 deadline itself.

Four practical conclusions.

Synthetic colours are the most avoidable additive category that exists. They provide no nutrition, no preservation and no safety function. Everything they do is visual. Whatever position you take on the behavioural evidence, there is no cost to avoiding them and no benefit forgone.

The behavioural evidence supports caution rather than alarm. The effect appears real for some children and is not uniform. If a child in your household seems sensitive, removing these dyes is a reasonable and low-cost experiment — and the EU's warning label exists precisely to enable that choice.

Regulatory status is not a reliable proxy for safety. Red 3 was legal in American food for thirty-five years after the same agency banned it from cosmetics. A substance being permitted may mean it was assessed and cleared, or it may mean no one has completed the process of removing it. Those are different situations that look identical from the shelf.

And the label is the only reliable check. Colour declarations are mandatory nearly everywhere, whether as E-numbers or FD&C names. Reading the last few lines of an ingredient list takes seconds and tells you more than any front-of-pack claim about what was added to make the product look the way it does.