Do Food Dyes Make Children Hyperactive?
The latest in a string of scientific misrepresentations by the Trump administration.
Artificial food dyes have been in the news a lot recently due to the Trump administration and the MAHA movement. There is a strong belief among these groups that artificial dyes are behind various problems that children face in the modern world, and that we can mostly fix children by banning the dyes. This was exemplified by a recent comment from FDA Commissioner Dr. Marty Makary:
In his quote, Dr. Makary references a study that supposedly proves that food dyes are bad. Specifically, he says “we have randomized controlled trial data in the Lancet”, which can only be a reference this famous 2007 randomized trial conducted in the UK. The study looked at whether various food additives caused children to be more hyperactive compared to a placebo.
The problem is that the data in this trial doesn’t really support the arguments that Dr. Makary is making. If anything, the study shows that artificial food additives probably have no impact on hyperactivity at all.
Let’s look at the science.
The Study
The study in question is a well-done piece of research that is fairly impressive for its time. Randomized trials have come a long way since the 00s, and its rare to come across papers that mostly still hold up 20 years later.
The research here was fairly simple. The authors took two groups of children, aged either 3 or 8/9, and gave them three different mixes of food additives - a high dose, lower dose, and placebo. All three were hidden in juice to disguise the placebo. The protocol was what’s known as a crossover design, so rather than having three different groups getting different substances, all of the kids got all of the mixes but at a random time so they didn’t know when they were getting additives vs placebo.
The additives tested in this study were many of the ones currently on trial in the United States, including Yellow 5 and 6, Red 40, and others. The children got either a 20mg or 30mg dose in the low and high dose groups, which the authors of the study said was equivalent to 2 or 4 bags of sugary treats.
The authors used three different measures to look at hyperactivity - parent, teacher, and study investigator ratings - and averaged these together to get a combined score. They then compared the children on this combined score when they were getting the food additives vs the placebo.
Given how the study is represented by Dr. Makary, you would have to assume that the data is stark. Impressive. Huge differences between the intervention and control, proving that the US government should take action immediately against these dangerous mixes.
Except that’s not what the study showed at all. What the authors found was a very mixed bag which is difficult to interpret. There was a small but statistically significant increase in hyperactivity for 3 year olds with the low dose mixture, but not the higher dose. There was a similar increase for 8/9 year olds with the high dose, but not the low dose. When the authors only looked at children whose parents said that they drank >85% of their juice and who had no missing data, the results were a bit stronger but still not conclusive.
These findings are very hard to understand if the food additives are really causing problems. For one thing, there’s no biological gradient. We expect bad things to have a gradient whereby low doses of the bad thing cause less of a problem than high doses. One bite of deathcap mushroom is bad - eating an entire beef Wellington contaminated with the deadly fungus is much worse.
But in this case, there was no gradient. In 3 year olds, those who got more additives had lower hyperactivity than those who got less. In 8/9 year olds, it was about the same. It’s very unlikely that these additives cause hyperactivity but a 50% increased dose has no additional effect.
The other problem is that the effect here is so small that it’s barely detectable. For those who care about statistics, the statistically significant p-values for the relationship between additives and hyperactivity were between 0.02 and 0.044. The standardized mean differences were 0.08-0.31. For those less statistically inclined, this means that while there were some increases in hyperactivity associated with the additives, they were very small and would be all but impossible to notice outside of a scientific study where such things are measured very precisely.
This was by far the biggest and best study on the question of food dyes for children. Most of the other research was done in tiny samples way back in the 70s and 80s, and is very hard to trust for many reasons. It’s not a definitive answer, but it’s the strongest one we’ve got right now.
Banning Dyes
Personally, I don’t have a strong opinion on whether food dyes should be banned in the United States. I come from a country where most of these dyes have never been approved - not because they’re unsafe, but because culturally no one has ever really wanted to use them - and our food is basically the same as anywhere else. In Australia, the only permitted colouring for breakfast cereal is Red 2, and even that is used quite sparingly. There are no artificial dyes approved for dairy products. You can used Red 3, but only in maraschino cherries and cake decorations. This is why repeated surveys by FSANZ, the Australia/New Zealand food authority, have found that rates of exposure to dyes for Aussie and Kiwi kids are extremely low.
So what Dr. Makary is proposing here is not unlike the current regulatory environment in Australia, and as someone who has traveled widely let me tell you that these dyes make little/no difference to food. Our Froot Loops are slightly less vibrant than the ones you can get in Boston, but it’s barely noticeable. Most artificial colours can quite easily be replaced with natural colourings that don’t impact the flavour of the product.
That being said, the evidence really does not stack up here. The biggest, best randomized trial that we’ve ever run looking at whether food additives make children hyperactive showed that they probably didn’t. That’s actually the conclusion of the FSANZ review and a similar review by EFSA, the European food safety agency. Even the FDA came to the conclusion that this study didn’t give any cause for alarm way back in 2011.
We also know that these food dyes almost certainly have no impact on hyperactivity in children more broadly because of Australia. If food colours were causing severe hyperactivity problems in children, we’d expect that the US would have much higher levels of ADHD for kids than Australia due to the much lower use of these artificial dyes. Instead, the rates of ADHD in both countries are about the same.
The most recent national data for Australia is from 2013/14 and indicates that 8.2% of children had ADHD at that time, compared to 9.8% of children in the US in 2016/2019. There are no current estimates for the prevalence of ADHD in Australia, but national drug prescription data shows that around 6% of kids aged 0-17 in the country are currently prescribed an ADHD medication which is similar to the CDC estimates of prescriptions in the US. Prescriptions in Australia are also currently skyrocketing, with a fourfold increase between 2013 and 2024.
Food additives simply can’t be the main explanation for this. It’s possible that they could contribute in some minor way, but since there hasn’t been any change in Australian food safety regulations regarding the colours in over a decade it’s hard to see how.
I don’t think banning food colours is a huge issue either way, but it is worth talking about because it’s yet another example of bad science being used by the Trump administration. Every rational scientific assessment has come to the conclusion that there’s no need to ban these dyes. The top piece of evidence cited by the FDA Commissioner as a reason to ban them has been independently reviewed by dozens of regulatory agencies globally and none of them have ever thought it shows a link between hyperactivity and the dyes.
It’s not a problem if everyone’s frosting becomes a little less vibrant. People will barely notice. But it is a problem that the people making decisions for the United States are doing so on the basis of obviously bad science.