At a White House event alongside the health secretary, the president gave pregnant patients a new thing to fear and physicians a new thing to say: "Effective immediately, the FDA will be notifying physicians that the use of acetaminophen during pregnancy can be associated with a very increased risk of autism." [1] The largest study ever conducted on exactly this question had already tested the idea, and the link it found dissolved the moment researchers controlled for the one thing that usually explains it.

What the biggest study actually found

The study, published in JAMA in 2024 by Ahlqvist and colleagues, followed 2,480,797 children born in Sweden over more than two decades. [2] In the simple version, comparing across the whole population, it did pick up a faint association between acetaminophen in pregnancy and autism, a hazard ratio of 1.05. [2] Then the researchers did the harder, more honest test: they compared siblings, children of the same mother, one exposed to acetaminophen in the womb and one not. Inside families, the association vanished. The hazard ratio for autism fell to 0.98, and for ADHD to 0.98, and for intellectual disability to 1.01, all statistically indistinguishable from no effect at all. [2]

The reason matters. When you compare unrelated families, a mother who took acetaminophen might differ in dozens of ways, genetics, chronic illness, the fevers that made her reach for the medicine, that also affect a child's development. Comparing siblings strips those shared factors out. When the signal disappears under that test, the most likely explanation is that the shared family factors, not the drug, drove the original correlation. An independent review in the Annals of Internal Medicine reached the same conclusion: no link to autism, ADHD, or intellectual disability. [3]

What the doctors say

The American College of Obstetricians and Gynecologists, the body that actually advises on pregnancy care, did not hedge. It called the suggestion that acetaminophen causes autism irresponsible, and pointed out the real-world stakes: acetaminophen is one of the only pain and fever relievers considered safe in pregnancy, and an untreated high fever carries documented risks to a developing fetus. [1] Frightening patients away from it does not remove a danger. It can introduce one.

Where the wording hides

The claim is built to survive a narrow fact-check. "Can be associated with" technically gestures at the weak, uncontrolled correlation that exists before the sibling comparison, while the framing, an FDA notice to physicians and a "very increased risk," tells parents to change their behavior as though causation were settled. It is not. Presenting a confounded correlation as a reason to act is how a null result gets dressed up as a warning.

THE BOTTOM LINE

  • "Very increased risk of autism" from acetaminophen in pregnancy: Mostly False [1]
  • Largest study, 2.48 million children: the association vanished with sibling controls (autism hazard ratio 0.98) [2]
  • ACOG calls the claim irresponsible; untreated fever in pregnancy carries its own risks [1]

A real question was asked, studied carefully, and answered: at the scale of millions of children, with the confounders removed, acetaminophen in pregnancy is not linked to autism. Telling physicians otherwise does not reflect the science. It overrides it.