US & Europe · Independent Updated August 2026
The Banned in Europe Registry
Explainers

The Southampton Study: The Six Dyes Behind Europe's Warning Label

One 2007 trial gave Europe a food-label warning America never adopted. Here's what it actually found, and the part both sides tend to skip: the effect was small, inconsistent, and came from a mixture nobody can pull apart.

By 8 min read Explainers

If you’ve ever picked up a bag of sweets in a London corner shop and noticed a small line of text warning that a colour “may have an adverse effect on activity and attention in children,” you’ve met the Southampton study. You just didn’t know its name.

That single sentence, printed on foods across the European Union, traces back to one trial run at the University of Southampton and published in 2007. It’s one of the cleanest examples of the transatlantic split this site keeps circling back to: same molecules, same science, two completely different answers on the label. Europe warns. America says nothing.

So it’s worth actually reading what the study found. Because the honest version is more interesting, and more useful, than either camp’s summary.

What the study actually did

The trial was commissioned by the UK’s Food Standards Agency and led by researchers at Southampton, with the lead paper published in The Lancet in 2007. It looked at two groups of ordinary children from the general population: 153 three-year-olds and 144 eight- and nine-year-olds. Not kids diagnosed with ADHD, not a clinical sample. Regular children, with the usual range of energy levels.

The design was genuinely good, which matters. It was randomised, double-blinded, and placebo-controlled, with a within-subject crossover. In plain terms: every child served as their own comparison, cycling through different drinks over six weeks, and neither the families nor the researchers scoring behaviour knew which drink a child was on at any given time. That setup strips out a lot of the wishful thinking that plagues diet studies, where parents who expect sugar to wind their kids up tend to see exactly that.

The children drank a fruit-juice mix each day. Some days it was a placebo. Other days it was one of two active mixes, each combining a preservative, sodium benzoate (E211), with a set of synthetic colours. Across the two mixes, the colours were tartrazine (yellow-5, E102), quinoline yellow (quinoline-yellow, E104), sunset yellow (E110), ponceau 4R (ponceau-4r, E124), allura red (red-dye-40, E129), and carmoisine (E122). The press quickly christened them “the Southampton Six.”

What it found, stated plainly

Averaged across the children, the active drinks nudged behaviour ratings toward more hyperactivity compared with the placebo. That’s the finding. It was statistically real in the pooled analysis, and it’s why the study made headlines.

Now the caveats, which are not footnotes. The effect was small. It wasn’t consistent across the two age groups, and it wasn’t consistent across the two mixtures. One mix showed an effect in the older kids but not clearly in the younger ones; the other split the opposite way. And crucially, every active drink combined several colours plus the preservative all at once. So even if you fully trust the behaviour scores, the study cannot tell you which additive did anything. It might have been one dye. It might have been the sodium benzoate. It might have been a combination that only misbehaves in a group. The design that made the trial rigorous, that all-at-once mixture, is the same thing that makes it impossible to convict any single ingredient.

That’s the tension at the heart of this. The study is real evidence of a small effect. It is not evidence that tartrazine, specifically, does anything to your child.

How Europe responded

Here’s where the two institutions in this story pull in different directions, and it’s the detail almost nobody knows.

Europe’s own food safety regulator looked at the Southampton data and was underwhelmed. In March 2008, the European Food Safety Authority’s panel on additives concluded the study offered “limited” evidence of a small effect on some children, and that because the additives were tested in mixtures, the results could not be used to change the acceptable daily intake for any of the colours or for sodium benzoate. Read that again: EFSA, the EU’s scientific body, declined to tighten a single limit on the basis of this trial.

The warning label didn’t come from the scientists. It came from the politicians. When the European Parliament was finalising a new food-additives regulation, members added the warning requirement late in the process, over and above what EFSA had recommended. It was written into Regulation (EC) No 1333/2008, and from July 2010 any food or drink sold in the EU containing one of the six colours has had to carry the line: the colour “may have an adverse effect on activity and attention in children.”

Notice what that label is, and isn’t. It’s not a ban. All six colours remain legal in Europe today. You can still buy products containing them; they just have to announce themselves. And the warning names the colours, not the sodium benzoate, even though the preservative was in every active mix in the study. The label is a political and precautionary instrument, a nudge aimed at manufacturers as much as parents. Faced with a choice between a warning line on the packet or reformulating with natural colours, most European brands quietly chose the latter. Which is a big part of why the sweets look different over there.

The twist the outrage stories leave out

If the story were “Europe followed the science and America ignored it,” you’d expect the EU’s limits to have gotten steadily stricter. They didn’t, quite.

In 2009, during separate re-evaluations, EFSA did lower the acceptable daily intake for three of the colours, quinoline yellow, sunset yellow, and ponceau 4R. Sounds like vindication. Except the reductions weren’t driven by the hyperactivity question at all; they came out of routine toxicology reviews. And then the pendulum swung back. When EFSA re-examined sunset yellow again in 2014 with better exposure data, it raised the acceptable daily intake back up, roughly to where it had started. Quinoline yellow’s exposure picture was reassessed too. The agency’s consistent position across all of it: the available data, the Southampton study included, don’t establish a causal link between any individual colour and behavioural effects.

So the tidy morality tale falls apart on inspection. Europe’s warning label is real and America has nothing like it, that part is true. But the label sits on top of a scientific verdict that is far more hedged than the label’s confident tone suggests. Europe warns and doesn’t ban. Its own scientists say the evidence can’t finger a culprit. Both things are true at once.

What this means when you’re standing in the aisle

America hasn’t been idle, for what it’s worth. The FDA has reviewed the dye-and-behaviour question more than once and concluded the evidence doesn’t support a causal link at the population level, though it acknowledges some children may be more sensitive than others. California went further than the feds, passing a 2024 law removing six dyes, including several of the Southampton colours, from food served in public schools. And the broader 2025 push to phase petroleum-based dyes out of the US supply has more momentum behind it than at any point in decades. The gap is narrowing, just unevenly and from the state level up.

If you have a kid who seems to bounce off the walls after a blue slushie, the reasonable read of the Southampton study is this: a real, small, hard-to-pin-down effect showed up in a careful trial, it doesn’t apply equally to every child, and the ingredient responsible was never isolated. That’s not a reason to panic. It’s a decent reason to notice that a fruit snack coloured with beet juice and one coloured with allura red are not obviously the same product, and to choose accordingly if it’s easy. Watching the total, the portions, the sleep, and the food culture around the sweets will do more for a kid than any single additive swap. The dyes are one variable in a much bigger picture, which is the same honest answer that runs through why so many people feel better eating in Europe.

The label on that London sweet bag isn’t proof of harm. It’s a country deciding that “we’re not sure, so we’ll tell you” is a reasonable thing to print. You can disagree with the policy and still find that a pretty defensible way to treat an adult holding a bag of sweets.

Sources