Gut Bacteria in Infancy and ADHD Risk: What the Study Found

Could the bacteria living in a baby's gut in the first week of life have something to do with an ADHD diagnosis ten years later? A newly published paper answers this question with "an association was seen." I should say right from the start that the answer is not "yes, this bacterium causes ADHD." There is an association, its cause is unknown, and there is no to-do list that follows from it.

I would like you to read this article in the same way: as an interesting scientific observation, as something that puts a burden on no one.

Who the study is from and where it comes from

This post is a summary of a paper published with open access in the journal Molecular Psychiatry. The paper was released under the Creative Commons Attribution 4.0 (CC BY 4.0) license. The findings belong to the authors of the paper; the account and the assessment here are mine.

The first author of the paper is Michael Widdowson and the last author is Jakob Stokholm. It is a study with twenty-two authors, and the team includes names such as Shiraz Shah, Jonathan Thorsen, Casper Sahl Poulsen, Bo Chawes, Klaus Bønnelykke, Thomas Werge and Bjørn Ebdrup. The paper was published in June 2026, in volume 31, issue 6 of the journal, on pages 3544-3557. I can put the title in plain words like this: Bifidobacterium in the newborn's gut is associated with the level of indole-3-lactic acid in the blood and with the risk of ADHD at age 10.

One note: I prepared this article not from the full text of the paper but from the abstract the authors wrote. Every finding here comes from that abstract. Where the abstract does not say something, I will say "the abstract does not say this," and I will keep general knowledge separate as general knowledge.

The type of study, in the abstract's words, is a birth cohort. A cohort means following the same group of people over time. The cohort used here is called COPSAC2010. The abstract describes this cohort as "deeply phenotyped," meaning that a large number of detailed measurements were collected about the children. The abstract does not say how many children there were or which country they were from. As general knowledge I know that COPSAC is a research group based in Copenhagen, and the author names fit that, but the abstract does not state it. The abstract does give age information at two points: the gut flora was taken from a sample collected when the baby was one week old, and ADHD was assessed at age 10.

What the study was asking

The abstract sets up its starting point in two sentences. The gut microbiome, that is, the community of microorganisms living in the gut, is said to be associated with brain health. It is experimentally supported that substances produced by microbes can affect the nervous system, and changes in gut flora have been linked to the mechanisms of some neuropsychiatric disorders.

On the ADHD side, according to the authors, most studies are cross-sectional, that is, studies that look at a single moment in time, and the newborn period has rarely been examined. I think this is a fair criticism. If you look at the gut flora of an adult or a child with ADHD and find a difference, you cannot tell whether the difference is a cause or a consequence. Measuring in the first days of life and looking at the outcome years later reduces part of this problem. But it does not remove it, and I will explain that below.

The authors' question is this: can the timing of first colonization and of particular bacteria, and the by-products they produce, influence the risk of developing ADHD later?

What was found

There are three findings, and all three fit into a few sentences of the abstract.

First: when the level of Bifidobacterium in the baby's one-week gut microbiome was higher, the association with ADHD at age 10 was stronger. The authors report this after genetic and early-life factors were taken into account "extensively." The abstract does not list which factors were taken into account. In statistics, "taking into account" means this: the model was adjusted for these factors so that an uneven spread of them between the groups being compared would not distort the result.

Second: indole-3-lactic acid, ILA for short, measured in the newborn's dried blood spot sample, mediates the association between Bifidobacterium and ADHD risk. ILA is a substance derived from an amino acid called tryptophan. Mediation is a statistical term: it means that part of the association of Bifidobacterium with ADHD seems to run through the level of ILA.

Third, and perhaps the most solid: the association between ILA in the newborn's dried blood spot and ADHD was replicated in two further independent cohorts. When another team finds a result pointing the same way in other children, that is more reliable than something seen in a single dataset. A subtle point to watch: the replication in the abstract is stated for ILA and ADHD. The abstract does not say that the association between Bifidobacterium and ADHD was also replicated.

As general knowledge: a dried blood spot is a small drop of blood taken from newborns for some screening tests. The abstract does not describe how this sample was collected.

The authors' interpretation, in their own words

The last part of the abstract is interpretation. The authors remind us that Bifidobacterium is known to support healthy neurodevelopment. They then say that the finding points to the timing of first colonization possibly being especially important for neurodevelopment. In particular, they suggest that a high level of Bifidobacterium-derived ILA in the first week of life may have negative consequences for the child's neurodevelopment.

This looks like an interesting contradiction. The same bacterium is known to be generally beneficial, and yet in this data its high level is associated with ADHD. The authors' explanation rests on "timing" rather than on "high": the problem may not be the presence of the bacterium but how much of it appears, and when, within the first week. The abstract does not offer a measurement that confirms this explanation; it leaves it as a hypothesis.

The last sentence of the abstract says that supporting an appropriate colonization pattern and ILA production could become clinical intervention strategies in the future. I am passing this on as the authors' claim, not as a recommendation by me or by the study. The study did not test an intervention; it observed. No one is being told to do anything, and I am not telling anyone either.

What this finding means for me

I am 29 and I got my diagnosis at 22. The children in this study are much younger than me, and I do not think this finding says anything about my own story. For me the study opens one more door while thinking about the causes of ADHD, and that is all.

I think the most important side of it is for parents. Someone whose child has ADHD may see this headline and think, "so something went wrong in the first week, was it my fault?" No such conclusion comes out of the abstract. The abstract does not say what determines the bacterial level in the first week, and it does not say that anyone did something wrong. I gathered what I know about the causes of ADHD in the article What causes ADHD?, and the share of heredity in Is ADHD hereditary?. As general knowledge: heredity is the best-documented factor in ADHD, and there is no single cause. This new finding can add, at most, one small piece to that picture. In the article Being a parent with ADHD I also made a bit more room for the feeling of guilt.

There is also the side of stigma. A biological finding, if read wrongly, can do harm in two directions: that ADHD is a "gut problem," or that it is "something preventable." Neither is what this study says. As I wrote in the stigma article, tying ADHD to a single cause often turns into a search for someone to blame.

What the study does not say

This is an observational birth cohort study. It does not prove causation. Even though extensive adjustment was made, some other factor that was not measured or not adjusted for may be affecting both the gut flora in the first week and ADHD later on. The abstract does not mention a method that would rule this out.

A mediation analysis also builds a chain of associations. To say that Bifidobacterium affects ILA and ADHD in sequence, a separate experiment showing that this chain really works that way in real life would be needed. There is no such experiment in the abstract. This is my reading: the mediation result should be read as "ILA may be a bridge," not as "ILA is the cause."

One study, one cohort. The ILA finding was replicated in two additional cohorts, which is reassuring. But the abstract does not give the size of those cohorts, where they were run, or the effect size. The effect size is not written anywhere, so we cannot learn from the abstract how strong the association is. Whether it is strong or slight, I do not know.

A pattern at group level does not describe an individual. The gut flora of a child in the first week does not by itself predict ADHD. The abstract does not offer a predictive tool either.

Finally, this is a proposal for a possible mechanism. In their own abstract the authors use cautious language such as "may" and "suggest." I am doing the same.

What stays with me

I like this study because it asks the question the other way round. Instead of linking ADHD to the gut years later, it starts with a measurement in the first week of life and follows the outcome over the years. I think a design like this, as the authors also point out, brings a time dimension that studies looking at a single moment miss. For me that is its main value.

But what the finding leaves in me is not excitement, it is caution. The connection between gut and brain has been talked about a lot in recent years, and topics like this quickly turn into "eat this, take that" advice. This study cannot be a basis for that. There is an association, the mechanism is only a possibility, and the study tried nothing. So I keep it as a curiosity of the day, not as a plan.

One more thing. Whenever findings like this come out, I ask the same question: could this information turn into a source of guilt for someone? If it can, then next to the explanation of the information it is necessary to write that "this is nobody's fault." There is no such claim in the abstract of this study. What is written is a bacterium, a substance and an association. The abstract does not explain why the association exists, so it does not offer any ground for blaming anyone.