What decades of research (and millions of children in Sweden and Japan) really tell us about acetaminophen and pregnancy.
It’s been a couple of weeks since Trump and RFK Junior told the world that acetaminophen (aka Tylenol or paracetamol) use in pregnancy causes Autism. Thankfully, there was considerable pushback from doctors and scientists worldwide.
I want to bring together the claims, the evidence, the studies and some overarching lessons to learn from this situation.
Background
Back in April this year, RFK Junior announced that they would know “by September” what was causing the “autism epidemic”. He said there would be “a massive testing and research effort” which would “involve hundreds of scientists from around the world.”
Just for some context, there are currently 33,171 studies and reviews published in scientific journals from 1957 onwards with the key words “autism” and cause” in the title. Tens of thousands of scientists have been studying this issue for decades.
The idea that the Trump administration would be able to find a definitive single ‘cause’ of autism in 5 months, where those 33,000+ studies have ‘failed’, was always a huge red flag.
But on the 22nd September, Trump delivered a speech where he told the world that they had discovered that acetaminophen use in pregnancy (the main ingredient in paracetamol and Tylenol) causes autism and told pregnant women not to take it and to “tough it out” if they’re in pain or have a fever. He said there were no “downsides” to this strategy.
He also announced they’d found a “cure” for autism, called Leucovarin.
So let’s take a look at these claims…
Acetaminophen
Trump’s claim wasn’t based on any new data or any new studies; it was based on a review paper that was published in August, which looked at 46 studies into acetaminophen use and various neurodevelopmental disorders (NDD), including Autism Spectrum Disorder (ASD) and Attention Deficit/Hyperactivity Disorder (ADHD).
The leading author behind this review was the Dean of the Harvard School of Public Health, Dr Andrea Baccarelli. While some used his credentials as the main reason why he was an authority on this subject, it has since come to light that he accepted $150,000 as an expert witness in a lawsuit against the makers of Tylenol in 2023, and that judges said that Dr. Baccarelli had “cherry-picked and misrepresented study results” in his testimony and was therefore “unreliable.”
Putting that potential conflict aside, what did the review show?
So the results were pretty mixed and basically confirmed what was already known on the subject – that there was some evidence of a possible association between acetaminophen use in pregnancy and a diagnosis of ASD, but the results were mixed, and there was certainly no causal link established.
The review concluded:
Our analysis demonstrated evidence consistent with an association between exposure to acetaminophen during pregnancy and offspring with NDDs, including ASD and ADHD, though observational limitations preclude definitive causation.
You could not conclude from these results that acetaminophen caused ASD or ADHD.
Despite this conclusion, Dr Marty Makary, the recently appointed head of the FDA, went on CNN and said:
“I’m reading now from a statement from the Dean of the Harvard School of Public Health – There is a causal relationship between prenatal acetaminophen use and Neurodevelopmental Disorders such as ADHD and Autism Spectrum Disorder”
This was not what the study showed, and it is dangerous for the head of FDA to go out there and claim there is a causal link when no such link has been found.
Confounders
At first glance, it can be easy to see a correlation or link between two factors and assume one causes the other.
But if that ever tempts you, you need to visit Tyler Vigan’s fantastic website on spurious correlations for some great examples:
It is obvious in these cases (I hope!) that there is no causal link between margarine consumption and divorces in Maine, or cheese consumption and bedsheet tanglings (800 deaths per year?!).
But it can be harder to accept that correlation doesn’t equal causation when the variables feel plausibly linked, as is the case with medicine use in pregnancy and neurodevelopmental disorders.
What we need to consider here is the reason why a pregnant mother will use Tylenol/Paracetamol. The advice has always been to use it sparingly and only if needed. Pregnant women aren’t taking it for fun!
They take it because they are in pain or they have a fever.
We also know that:
- Having an infection during pregnancy is a known risk factor for ASD, as is having a fever.
- Maternal obesity is also a known risk factor for ASD, and obesity is associated with higher rates of chronic pain.
- The biggest risk factor associated with being diagnosed with ASD is genetics – having a parent with ASD. And there’s evidence showing that those with ASD can be more sensitive to pain, and could therefore be more likely to need pain relief.
These are all confounding factors that may explain the association.
It may not be the acetaminophen causing the ASD, but the infection or fever that made the mum take it.
Or the obesity in the mum that caused the chronic pain, which made her take it.
Or the fact that mum is also autistic and has hypersensitivity to pain, and therefore is more likely to need pain medication in pregnancy…
What about the Swedish study?
This is where the “Swedish study” comes in.
This study examined the health outcomes of ~2.5 million children born in Sweden between 1995 and 2019, linking them to their mothers’ records to investigate prenatal exposures and subsequent health outcomes in childhood.
The risks of developing ASD or ADHD when looking at the population as a whole were very slightly higher in those who had been exposed to acetaminophen in pregnancy.
But when they looked at siblings and compared when the mum had taken acetaminophen in one pregnancy and not the other, this association vanished. The sibling exposed to acetaminophen in pregnancy was no more likely to be diagnosed with ASD or ADHD than the sibling not exposed.
By removing many of the confounding variables such as maternal health, weight, genetics, family environment, education, income, etc, the link between acetaminophen and ADHD and ASD disappeared.
Other sibling studies
A similar method was used in a Norwegian study in 2021, which found no association between short-term acetaminophen use in pregnancy and ADHD in the child when looking at sibling pairs.
A more recent study (Sep 2025) from Japan wanted to look at whether these findings would be replicated in an Asian population. They looked at over 200,000 children and found a slight increase in risk of ADHD and ASD overall when exposed to acetaminophen, but again, that association disappeared when comparing siblings.
These very large studies, designed specifically to answer the acetaminophen ASD/ADHD question (that took place long before Trump and RFK claimed to be the only ones who cared to ask the question!), give us pretty solid evidence that there is no causal link.
Surely there’s no harm in avoiding it just to be safe?
That was the message Trump gave. That there is no downside to not taking it, and pregnant women should just “tough it out”.
But we’ve seen that the reasons why acetaminophen might be needed are already known risk factors, not just of ASD but of premature birth, neural tube defects and congenital heart defects.
The Harvard Study included this warning in their conclusion:
While this association warrants caution, untreated maternal fever and pain pose risks such as neural tube defects and preterm birth, necessitating a balanced approach. We recommend judicious acetaminophen use—lowest effective dose, shortest duration—under medical guidance, tailored to individual risk–benefit assessments, rather than a broad limitation.
Despite this being in the study Trump was using, he ignored that warning, instead telling pregnant women to “tough it out” when experiencing symptoms that are known to increase the risk of a whole host of medical issues in babies, to lower the possible theoretical risk of ASD, when the evidence to date does not suggest a causal link…
If a pregnant woman is suffering from pain or fever, not taking acetaminophen (the only medication that is considered safe in pregnancy) is not a zero-risk strategy.
In fact, this study compared mothers with a fever during pregnancy and found that those who took medication (like acetaminophen) had a 30% higher odds of their child being diagnosed with autism than mothers with no fever. But those who had a fever and didn’t take medication had more than double the odds.
To put those study results into easy-to-understand numbers:
- For every 1000 babies where there is no fever in pregnancy, ~ 10 may develop ASD
- For every 1000 babies, where there is a treated fever in pregnancy, this could increase to ~ 13
- For every 1000 babies, where there is an untreated fever in pregnancy, it could increase to ~ 26
In other words, treating the fever reduced the risk substantially.
Leucovorin
Alongside the announcement that they had found the cause of autism, they also declared they had found a treatment for it.
Trump has decided that the FDA are going to approve the drug Leucovorin as a treatment for autism. This shocked the scientific community, as there have only been a handful of trials looking at Leucovorin to treat ASD, with only two placebo-controlled trials containing 67 children in total (across the two trials).
Drugs are never given FDA approval without large, double blind, placebo-controlled trials with thousands of participants. This is absolutely unprecedented and seriously highlights the undermining of scientific standards happening in USA right now.
It may be that Leucovorin does turn out to be a wonder drug that ‘cures’ Autism, but we do not have the evidence to show that, and it is a dangerous precedent being set, allowing politics to decide what drugs will and won’t be approved, without rigorous scientific testing.
A cautionary tale
β-carotene provides us with a cautionary tale which highlights both the dangers of assuming correlation equals causation and of reaching conclusions without robust, large-scale evidence.
Back in the 1970s and 1980s, observational studies showed a correlation between diets that were high in β-carotene and a decreased risk of lung cancer.
This led to a booming industry in β-carotene supplements, with people believing they would be protected from lung and other cancers if they took them
When large, randomised, placebo-controlled trials were carried out in the early 90s, they discovered that smokers taking β-carotene supplements were actually at higher risk of lung cancer.
The early observational trials had just highlighted that those who ate a good, healthy diet (which happened to contain higher levels of β-carotene) were also living healthier lifestyles in general and therefore had a lower risk of developing lung cancer. It was only these rigorous trials of β-carotene supplements that highlighted the increased risks of taking a β-carotene supplement for smokers.
The correlation between β-carotene and lower lung cancer risk was just that – it was a correlation. Eating a healthy diet with lots of fruits and vegetables was the third factor that tied the variables together.
Neurodiversity
I want to end with highlighting that Autism Spectrum Disorder is not a ‘disease’ that is caused by a singular factor, or that can be ‘cured’ by a single drug. It is a developmental difference affecting many aspects of the brain, involving thousands of genes covering socialising, speech, learning, and behaviour.
There is a reason why it is 2025, and after more than 70 years and 33,000+ studies looking at the causes of autism, no single answer has been found! (And no, it’s not because scientists and doctors are covering it up!)
And while some people on that spectrum will need lifelong support, for whom targeted therapies to ease certain symptoms would be a lifeline, there are also millions of people for whom neurodiversity is not something that is broken to be repaired, but is part of who they are. They can bring different ways of thinking and seeing the world that enrich society, science, art, music, and so much more.
That’s why framing autism or ADHD as conditions that must be “prevented at all costs” or “cured” is misleading and harmful. It risks reviving old stigmas, rather than recognising both the challenges some autistic people face and the value that neurodiversity brings.
Thank you for sticking with me through another data-heavy topic. Please do share it with anyone you know who has been concerned by the Tylenol/Acetaminophen announcement.
The anti-science narratives coming out of the White House at the moment are terrifying and seriously damaging not just to the health and well-being of millions, but also damaging trust in the doctors and scientists who have dedicated their lives to answering the sorts of questions they’re now being accused of covering up.
This article was first published in Emma’s Substack, MonkDebunks and is reproduced here by kind permission of the author.