24/09/2025
Paracetamol doesn't cause autism - let me explain why the research doesn't stack up. 🤰💊
I have seen well-meaning advocacy announcements saying that "there isn't sufficient evidence" that there is a link, or that "it isn't proven."
These statements can make us insert the word "yet" at the end... Because they make it sound like we will find a link if we keep looking. That where there is smoke, there is fire. Surely there must be something in it (this is why the vaccine–autism link is so sticky even though it's been 100% scientifically debunked).
Methodological critique of autism research is my day job and so maybe I can help explain instead of just saying "take my word for it, it's a lot of crap" (which is what most experts have implied).
📚🔍In the meta-analysis (summary of 47 studies) RFKJn and Trump are alluding to, a weak (5-7%) association was found between taking the over-the-counter pain killer acetaminophen (sold as Paracetamol or Tylenol) in pregnancy and having a child diagnosed as autistic. But there are confounders, a fancy way of saying there are other reasons for the (weak) link.
(1) Mothers experiencing infection, fever, migraine, or pain may have other conditions (eg, diabetes, Ehlers-Danlos Syndrome) that we already know are related to autism.⚖️🩺
(2) We know autism is highly heritable. Autistic mums or mums with autistic traits have greater sensitivity to pain and different experiences of pain due to sensory sensitivities. 🧬🤰https://tinyurl.com/a46mt5z2
In both cases, this may mean that women may be more likely to use pain killers in pregnancy. They pass on their genes (related to autism) to their children who – boom – are more likely to be diagnosed as autistic.
(3) Recall biases. It is well know that people look for a reason for why they or their child have a condition. When asked, they look back and are more likely to remember doing things that they suspect may have had an effect (like taking medication).
This is often seen in observational research. For example, it was long thought that a bump to the breast could cause breast cancer because women with breast cancer were more likely to remember and report receiving a bump.
Notably, the authors of the meta-analysis themselves conclude :“We cannot answer the question about causation". And on the day of Trump's presser on the issue, the FDA said “a causal relationship has not been established” https://tinyurl.com/y62hh5tw.
(4) There is robust evidence from a recent population-based sibling-controlled study published in the highly reputable Journal JAMA in 2024 involving 2.5 million children.
Mothers' use (and amount) of pain killers in pregnancy was compared for their autistic children with their non-autistic children. This controlled for "confounders" as the mother's pain sensitivity, medical conditions, and environment were basically the same for both children.
The study measured use of pain killers during pregnancy prospectively – that is, tecorded from antenatal and prescription records during pregnancy rather than from remembering later). So there were no recall biases. This is a much more scientific way to determine causality.
📣The study concluded that there was "no association with autism". And they suggest associations observed in other models may have been attributable to confounding (see 1-3). https://tinyurl.com/2han37wb
🩺⚖️We know paracetamol should be used sparingly in pregnancy, and is toxic in overdose. But in cases of high fever, it is very dangerous for the mother and the developing baby if it isn't brought down by this medication.
And you know what? Autistic people reflect natural human diversity and bring so many gifts to our world. They have been in human history forever and have give us some of the world's finest thinkers, inventors, and artists. Let's not talk about autism as if it's a disease to be prevented or cured. 🙏❤️♾️ 🧠
What would I know?
I am a systematic reviewer specialising in autism, and have published 17 systematic reviews in the area. I deliver the Living Guideline programme of work for the NZ government and have led annual revisions to the NZ Autism Guideline since 2009, and revised its 2nd and 3rd editions.
