How many times does a link need to be disproven before it stops being treated as an open question? The latest answer comes from Hong Kong, where researchers analyzed more than 700,000 mother-child pairs and found, once again, no association between acetaminophen use during pregnancy and autism or ADHD in children.
The study, published June 29 in JAMA Internal Medicine by a team led by Eric Yuk Fai Wan of the University of Hong Kong and Ian Chi-Kei Wong of Aston University, is the third sibling-matched cohort study on this question since 2024 and the largest conducted in an Asian population. Its conclusion is unambiguous: prenatal acetaminophen exposure, whether measured by trimester, dose, or duration, showed no increased risk for either neurodevelopmental condition.
What the study actually did
The researchers drew on electronic health records from Hong Kong spanning 2001 to 2023, covering more than 700,000 mother-child pairs. Approximately 43% of children, roughly 300,000, had documented prenatal exposure to acetaminophen (known as paracetamol outside North America).
The key design feature was sibling matching. By comparing siblings born to the same mother, where one was exposed and the other was not, the study controls for shared genetic and environmental factors that have confounded earlier research. The autism sub-analysis included more than 124,000 children from discordant sibling pairs; the ADHD analysis included more than 97,000.
The lack of association held across every subgroup examined, regardless of which trimester the drug was prescribed, the dose, or the frequency of use. The researchers also ran negative control analyses looking at paternal prescriptions and maternal prescriptions after pregnancy, which returned similarly null results. These controls are important: if earlier positive signals had been real, one would not expect to see null results from paternal exposure or from the same mother after delivery.
The accumulating evidence
This study joins a converging body of evidence that has grown increasingly difficult to dismiss. The largest study to date remains the Swedish nationwide sibling analysis published in JAMA in 2024, covering approximately 2.48 million children, which found no association (autism hazard ratio 0.98, 95% CI 0.93 to 1.04; ADHD hazard ratio 0.98, 95% CI 0.94 to 1.02). A Japanese nationwide study of 217,602 children, published in 2025, similarly found no link. A Taiwanese analysis of roughly 2.09 million births, published this year in JAMA Pediatrics, reported an autism hazard ratio of 0.98 (95% CI 0.90 to 1.07) and an ADHD hazard ratio of 0.99 (95% CI 0.96 to 1.03). A Norwegian sibling study from 2021 also came up null.
Taken together, these five sibling-controlled studies span Sweden, Norway, Japan, Taiwan, and now Hong Kong, covering millions of children across diverse populations and healthcare systems. Every one of them reached the same conclusion.
In November 2025, a BMJ umbrella review by Sheikh and colleagues that examined nine systematic reviews encompassing 40 individual studies concluded that sibling-controlled analyses show no association and that earlier positive signals were attributable to confounding, specifically, that the underlying conditions for which acetaminophen was taken (fever, infection, pain) were themselves independently associated with neurodevelopmental outcomes.
Why the question persists
Despite this accumulation of evidence, acetaminophen’s safety in pregnancy remains a topic of public concern and litigation. In December 2023, a federal judge in the Tylenol autism and ADHD mass tort litigation barred expert witnesses from testifying for the plaintiffs, ruling that their claims lacked scientific support. In September 2025, the White House held a briefing impugning acetaminophen’s safety and linking it to autism, without evidence, as Nature subsequently fact-checked. The American College of Obstetricians and Gynecologists continues to recommend acetaminophen as the “first and safest choice for fever or pain relief during pregnancy.”
The accompanying editorial in JAMA Internal Medicine, by Brian K. Lee of Drexel University and Viktor H. Ahlqvist of the Karolinska Institute (both authors of the 2024 Swedish study), is titled “Acetaminophen Use in Pregnancy and Neurodevelopmental Outcomes, Reassuring Evidence From a Matched-Sibling Study.” The title says what needs to be said.
Caveats
The study has limitations. It relied on prescription records, meaning over-the-counter acetaminophen use, common and largely unmeasured, was not captured. However, this would be expected to bias toward the null only if OTC use differed systematically from prescribed use. Sibling designs control for shared familial factors but cannot fully rule out confounding by differences between individual pregnancies, such as a severe fever in one pregnancy but not another. The confidence intervals, while narrow, cannot definitively rule out minuscule absolute risk differences that would be clinically meaningless.
These limitations, however, are shared by every observational study on this question. The consistent null result across five independent sibling-controlled studies, a BMJ umbrella review, and millions of children makes it difficult to argue that the question remains scientifically open.
At a certain point, the evidence is what it is. Asking how many times a link needs to be disproven is not a rhetorical question, it is one that the data have now answered five times over.
Source: Luo S, Gong Q, Ai Y, et al. Prenatal Acetaminophen (Paracetamol) Use and the Risk of Autism and/or ADHD Among Sibling-Matched Cohorts. JAMA Internal Medicine. Published online June 29, 2026. doi:10.1001/jamainternmed.2026.2215