Food Dyes and Hyperactivity: What the Southampton, FDA, and OEHHA Reviews Actually Found

The short answer

Three reviews, three different conclusions, and none of them are wrong exactly — they asked slightly different questions and set the evidentiary bar in different places. The 2007 Southampton study found an effect. The FDA's 2011 committee looked at that and more evidence and declined to act on it for children generally. California's OEHHA looked again in 2021, with a decade more research available, and concluded the effect was real enough that federal safety limits should be revisited. This page walks through what each one actually did, rather than which side of a debate to pick.

The study that started this round: Southampton, 2007

In 2007, a University of Southampton team led by Donna McCann published a randomized, double-blind, placebo-controlled trial in The Lancet — the kind of study design that carries real weight, because neither the families nor the researchers scoring behavior knew which drink a child had been given.

The trial included 153 three-year-olds and 144 eight- and nine-year-olds, a general-population sample rather than children pre-selected for behavioral problems. Over several weeks, children drank a daily juice drink containing either a placebo or one of two mixtures of synthetic dyes — combinations of tartrazine (Yellow 5), sunset yellow (Yellow 6), carmoisine, ponceau 4R, quinoline yellow, and allura red (Red 40) — plus sodium benzoate, a common preservative. All the drinks looked and tasted identical.

The result: both dye-and-benzoate mixtures increased hyperactive behavior, measured by a combined score from parent and teacher ratings and, for the older children, a computerized attention test, relative to placebo. The effect showed up in both age groups and across the general population sample — not only in children already flagged as hyperactive.

That finding had real regulatory consequences, just not in the US. The European Union responded by requiring a warning label on foods containing any of the six dyes in the study: a statement that the product "may have an adverse effect on activity and attention in children." That label requirement is why European versions of some products use different colorants than their US counterparts to this day.

The FDA's 2011 review: evidence acknowledged, action declined

The Southampton findings, along with a formal petition from the Center for Science in the Public Interest, led the FDA to convene its Food Advisory Committee in March 2011 specifically to evaluate whether synthetic color additives should carry a warning label or face further restriction in the US.

The committee reviewed the Southampton trial along with the broader literature, including earlier and smaller elimination-diet studies going back to the 1970s. Its conclusion, in the FDA's own language, was that a causal relationship between exposure to color additives and hyperactivity in children in the general population has not been established — but that for certain susceptible children, including those with ADHD or related behavioral problems, existing data suggest their condition may be exacerbated by a number of substances in food, color additives among them.

That is a narrower finding than "dyes are safe," and a narrower finding than "dyes cause hyperactivity." It says: not proven for most children, plausible for some. The committee then voted, 8 to 6, against recommending a warning label or further restrictions — a close vote on a genuinely close question, not a unanimous dismissal. The FDA's current consumer-facing position, still published today, keeps that same shape: most children show no adverse effects from color additives, but some evidence suggests certain children may be more sensitive.

California OEHHA's 2021 review: a decade of new evidence, a different conclusion

California's Office of Environmental Health Hazard Assessment (OEHHA) revisited the question a decade later, publishing its findings in April 2021 after a two-year evaluation covering seven FDA-approved synthetic dyes, including Red 40 and Yellow 5. Unlike the FDA's 2011 review, it had roughly ten additional years of published research to draw on, including newer human trials and animal studies looking at brain chemistry and structure, not only behavior scores.

OEHHA's conclusion went further than the FDA's: that synthetic food dyes can cause hyperactivity and other neurobehavioral problems in some children, that sensitivity to dyes varies meaningfully from child to child, and — the part with the most practical teeth — that the federal intake levels considered safe were set using studies conducted 35 to 70 years ago and may not adequately protect children's behavioral health by current evidence. The report also flagged that typical daily dye exposure for many children, mostly from juice and soft drinks, already runs close to or above some of those older safety thresholds for at least one dye.

It's worth being precise about what that report is and isn't. It's a state risk assessment, produced to inform California policy discussions — it did not change what's legal to manufacture or sell anywhere, including in California. The seven dyes it reviewed remain FDA-approved nationwide. What it did was put a more emphatic scientific conclusion on the record than the FDA had reached ten years earlier, using more evidence than the FDA had available at the time.

Why these don't actually contradict each other as much as they look like they do

Read side by side, "not established" (FDA, 2011) and "can affect neurobehavior" (OEHHA, 2021) sound like opposite verdicts. They are not, and the reason is a measurement problem this site covers in general terms on the shady truth about artificial food dyes: a strong effect in a minority, averaged together with no effect in the majority, produces a weak group-level number — which is what a trial reports. Rather than restate that here, this page uses it to do the thing only this page can: line the two reviews up and show that they were not answering the same question.

FDA's 2011 committee was applying a higher bar — proof of a causal effect across the general population — and didn't find it met. OEHHA's 2021 review was asking a different, narrower question — can dyes affect some children — and concluded the evidence supported yes, with the added finding that current safety margins may not be generous enough for those children. Both conclusions can be correct about the population they're each describing. The genuine disagreement is about what a subset-level effect like this one should obligate a regulator to do — nothing, a warning label, or a lower safety threshold — and that's a policy judgment as much as a scientific one, which is exactly why the US and EU landed in different places on the same underlying research.

What this means if you're deciding what to do about it

There is no blood test or screening tool that tells you in advance whether your child is one of the sensitive ones. The only individual answer available is a careful elimination: remove the dyes properly for two to three weeks, track behavior, then reintroduce and watch again. If nothing changes on either end, that's a real answer — you can stop spending attention on it. If something does change both ways, you have a more useful result for your own child than any study average could have given you.

Whichever direction you land on, the dyes involved in the Southampton mixtures — Yellow 5, Yellow 6, Red 40, and the others — are the same six dyes currently sitting in the FDA and HHS's voluntary phase-out request, not the one dye (Red 3) under a binding deadline. See what that phase-out actually requires if you want the regulatory picture, or start with where Red 40 actually shows up if you're ready to run the elimination and need to know what to pull first.