Does Tylenol in pregnancy cause autism? Weighing the evidence a year after the White House warning

In September 2025 President Trump told pregnant women not to take Tylenol, linking the painkiller to autism, and the FDA moved to change the drug's labelling. A year on, the claim is still repeated widely, and new studies keep landing. The question on trial: does the evidence show that acetaminophen (called paracetamol outside North America) taken in pregnancy causes autism?

Background

Acetaminophen, the active ingredient in Tylenol, has long been the standard recommendation for pain and fever in pregnancy, partly because the alternatives carry known risks. For years some observational studies have found that children whose mothers reported using it were slightly more likely to be diagnosed with autism or ADHD. The central scientific dispute is whether that association reflects the drug or other factors, such as the illnesses that lead mothers to take it and the genetics families share.

On Sept. 22, 2025, Trump and Health Secretary Robert F. Kennedy Jr. announced autism initiatives that included the acetaminophen warning (HHS press release). The FDA said it was starting a label change and sent a notice to physicians. In its own announcement, the FDA said that "a causal relationship has not been established" and that there are contrary studies. Reuters reported that the manufacturer and many medical groups said the evidence did not support a causal link (Reuters).

Where things stand (as of Oct 18, 2026)

• The strongest evidence cited by the administration's side was a 2025 review in Environmental Health by researchers including Harvard and Mount Sinai scientists. It concluded the evidence was consistent with an association and recommended judicious, medically guided use rather than avoidance.
• Studies that compare siblings (one exposed in the womb, one not) help control for shared genes and family environment. A large 2024 Swedish sibling study in JAMA found that associations seen in crude data disappeared in sibling comparisons.
• In January 2026 a systematic review and meta-analysis in The Lancet Obstetrics, Gynaecology and Women's Health reported no association with autism, ADHD or intellectual disability in sibling-comparison analyses.
• In April 2026 a Danish nationwide study of about 1.5 million children found no significant association between prescribed acetaminophen in pregnancy and autism, including in sibling-matched analysis (Reuters, CNN). A sibling-matched cohort from Hong Kong published in JAMA Internal Medicine also reported no increased risk.
• Medical guidance has not changed in substance: use acetaminophen when needed, at the lowest effective dose for the shortest time (Reuters explainer). Untreated fever in pregnancy carries its own risks.

Key claims on trial

• "Taking Tylenol in pregnancy causes autism." (Trump administration, some advocates). Unverified. No study has established causation, and the FDA itself said so. The best-controlled recent studies (sibling comparisons in Sweden, Denmark and Hong Kong, and the 2026 Lancet meta-analysis) find no link.
• "Studies show an association between prenatal acetaminophen and autism or ADHD." (administration, some researchers). Verified, with caveats. Some observational studies do report associations, as the 2025 Environmental Health review found. The disputed part is what explains them. Associations tend to shrink or vanish when siblings are compared.
• "The FDA acted on new evidence." (administration). Disputed. The FDA notice relied on existing observational literature, including the 2025 review. It did not cite a new trial. Supporters say the accumulated literature justified caution regardless.
• "Sibling studies settle the question." (critics of the claim). Disputed. Sibling designs control for shared genes and environment, but they have limits of their own, including how exposure is measured and how many discordant sibling pairs exist. The Danish study covered only prescribed acetaminophen, not over-the-counter use. Several sibling studies now point the same way, which strengthens the conclusion without making it absolute.
• "Acetaminophen remains the safest option for fever and pain in pregnancy." (medical groups). Verified as current guidance. This is the position reported by Reuters for major medical groups and is consistent with the FDA notice, which did not ban or restrict the drug.
• "Pregnant women should avoid Tylenol entirely." (Trump, Sept. 2025). Disputed. Even the review the administration relied on recommended judicious use under medical guidance, not avoidance. No regulator has advised a ban.

What is still unknown

Whether very high or prolonged doses carry a small risk that current studies cannot detect. How the FDA label wording will ultimately read, and whether it will be revised in light of the 2026 studies. Whether public messaging has changed how pregnant women treat fever, which researchers say should be tracked. No randomised trial exists, and none is likely for ethical reasons, so the question will continue to rest on observational data.

Source notes

The HHS and FDA announcements come from agencies that are parties to the dispute, though the FDA text is notably more cautious than the President's remarks. Kenvue, the maker of Tylenol, has a commercial interest in the drug's reputation, and its statements are not relied on here. One senior author of the 2025 Environmental Health review, Harvard public health dean Andrea Baccarelli, disclosed in the paper that he served as a plaintiffs' expert witness in acetaminophen litigation, and the New York Times reported he was paid about $150,000 for that work. A federal judge had earlier excluded his proposed testimony in that litigation as unreliable (court opinion). This does not by itself make the review wrong, but readers should know it. Details of the Danish study were confirmed from Reuters and CNN search excerpts and an obstetrics trade report rather than the full journal paper. The September 2026 umbrella review in Developmental Medicine and Child Neurology was seen only in a search excerpt and is listed for further reading, not relied on.

Sources

• FDA announcement, Sept. 2025 (party source)
• FDA notice to physicians
• HHS press release (party source)
• Reuters: Trump links autism to Tylenol and vaccines
• Reuters explainer: Is Tylenol safe in pregnancy?
• Prada et al., Environmental Health (2025)
• Ahlqvist et al., JAMA (2024)
• Lancet Obstetrics, Gynaecology and Women's Health meta-analysis (2026)
• JAMA Internal Medicine, Hong Kong sibling cohort (2026)
• Reuters: Danish study finds no link (Apr. 2026)
• CNN: Danish study
• New York Times: Harvard dean paid as expert witness in Tylenol lawsuits
• US District Court (SDNY) Daubert opinion in acetaminophen litigation
• Umbrella review, Developmental Medicine and Child Neurology (search excerpt only)

Lead image: President Donald Trump announces findings on autism in the Roosevelt Room of the White House, Sept. 22, 2025. Photo: Joyce N. Boghosian/The White House, via The White House.