Is Trump right to link autism with Tylenol?

Donald Trump’s apparent suggestion that people could protect themselves against Covid by injecting themselves with bleach marked a low point in his first administration. It provided his critics with evidence that he was an erratic president trying to ride roughshod over scientific evidence as well as common sense. It is easy, therefore, to dismiss the American president’s announcement that government health warnings will henceforth be printed on packets of Tylenol – the brand name for acetaminophen – telling pregnant women to avoid the painkiller for fear it will cause autism in their unborn children as yet another anti-scientific diatribe.

The involvement of health secretary Robert F. Kennedy Jr – a long-term vaccine skeptic – adds to the impression that the association between autism and acetaminophen might be a little cooked-up. The American College of Obstetricians and Gynecologists lost no time in branding the presidential announcement as “irresponsible”.

But is there any genuine link between autism and the consumption of Tylenol? There is quite a lot of evidence on this and interestingly, it doesn’t entirely dismiss a link, although if there is one, it does not appear to be very strong.

A review of the evidence was published in the journal Environmental Health in August – carried out by a team of scientists from several universities, including Harvard and the University of California. It looked at 46 studies, 27 of which found a link between acetaminophen use and neurodevelopmental disorders in children (not just autism but also attention deficit and hyperactivity disorder, ADHD). Of the others, nine found a null link and four found a negative association – i.e., suggesting that acetaminophen could actually lower the risk of neurodevelopment disorders. It didn’t classify the remainder of the studies into either of those groups. Pointedly, however, the review suggested that the higher-quality studies were more likely to show a positive association between acetaminophen use and neurodevelopmental disorders.

But how big is the link? One of the most comprehensive studies on this subject uses data on 2.5 million Swedish children born between 1995 and 2019. It found that 1.42 per cent of children whose mothers had taken acetaminophen during pregnancy went on to develop autism, compared with 1.33 per cent of children whose mothers didn’t take the painkiller. There are other things to consider behind this rather weak association – mothers who took acetaminophen were quite likely to have been in worse general health than those who did not, so their acetaminophen use is surely not the only thing going on here.    

Yesterday’s announcement is not purely some off-the-cuff move by Trump – it is backed by Jay Bhattacharya, Director of the National Institutes of Health (whose background is nevertheless in economics rather than medicine). He was one of the signatories of the Great Barrington Declaration in 2020, which called for young people less at risk of Covid to be allowed to get on with their lives during the pandemic.

While evidence for any link between Tylenol and autism is certainly not strong, it is not unreasonable to ask whether pregnant women – and many other people, for that matter – should try to avoid taking Tylenol if they can. Taking medical drugs is often a trade-off between risk and reward, and while the risks in this case might not be great, nor, in many cases, will be the rewards.

A lot of people are taking painkillers far too routinely without considering that pain is there for a reason: it is telling you not to put too much weight on that injured ankle or warning you that there might be some serious problem in your stomach. Kill the pain and you kill the warning with it.       

The presentation of the Trump administration’s policy, however, is dreadful. Trump’s assertion that the Amish community don’t have autism because they don’t take painkillers does seem a little dubious, as does Robert F. Kennedy’s claim that there aren’t many 70 year olds with full-blown autism. The diagnosis of autism has certainly increased dramatically in recent decades but it seems to me to be strongly related to it being a fashionable diagnosis. There are plenty of 70 year olds living in institutions who were never diagnosed with autism when they were young but who would be now.