European review finds no causal link between prenatal paracetamol and autism, ADHD

A comprehensive European systematic review and umbrella analysis published in The Lancet Obstetrics, Gynaecology & Women's Health finds that the best-quality available evidence does not support a causal relationship between maternal paracetamol (acetaminophen/Tylenol) use during pregnancy and offspring neurodevelopmental disorders, including autism spectrum disorder and attention-deficit/hyperactivity disorder.

The paper pooled and re-examined multiple observational studies that had reported positive associations between prenatal paracetamol exposure and child neurodevelopmental outcomes. The authors report that apparent links largely weakened or disappeared after more rigorous control for confounding factors and when analyses prioritized higher-quality evidence. The reanalysis was prompted in part by U.S. health officials’ recent statements and by a contrasting 2025 meta-analysis co‑authored by Andrea Baccarelli that reported “significant links.”

Previous literature has been mixed. Multiple earlier reviews, summarized in a BMJ analysis, consistently reported small positive associations, and four meta-analyses produced pooled risk estimates for ADHD in the 1.2–1.4 range with smaller estimates for autism. Subgroup and sensitivity analyses in some studies showed stronger associations with longer duration or higher frequency of paracetamol use and with third-trimester exposure. Yet seven of nine reviews cited in those summaries warned explicitly against interpreting observational associations as causal because of study limitations, bias, and unmeasured confounding.

More rigorous study designs have weakened the case for causation. An NIH-funded sibling-comparison study publicized in an April 11, 2024 media advisory used the Swedish Medical Birth Register and Swedish Prescribed Drug Register to compare siblings with differing prenatal acetaminophen exposure. That analysis found an apparent small increase in risk before sibling comparisons that was no longer evident when controlling for shared familial factors, suggesting that genetic, environmental, or parental health factors likely explain previously reported associations rather than a direct drug effect.

Methodological problems recur across the literature. Commentaries and reviews, including analyses by FIGO and BMJ, highlight possible recall bias, inconsistent or missing information on dosage and duration, heterogeneous outcome assessments, and inadequate control for genetic and environmental confounders. These limitations reduce the ability of observational studies to support causal inference.

The debate has also included questions about conflicts of interest. FIGO notes that Andrea Baccarelli served as a paid expert witness in class-action litigation against paracetamol manufacturers in 2023, and that his testimony was reportedly rejected by the court as scientifically unfounded.

Regulators have taken cautious, measured positions. The UK Medicines and Healthcare products Regulatory Agency said, “there is no evidence that taking paracetamol during pregnancy causes autism in children.” The European Medicines Agency confirmed in September 2025 that current recommendations remain unchanged. The U.S. Food and Drug Administration, which initiated a label-change process in September 2025, stressed that “while an association between acetaminophen and neurological conditions has been described in many studies, a causal relationship has not been established.”

Clinicians and pregnant patients must balance established clinical considerations: acetaminophen remains widely used in pregnancy because alternatives such as NSAIDs and opioids carry known fetal risks, and treating fever can reduce risks of birth defects. The Lancet analysis, the NIH sibling evidence, and longstanding methodological critiques together argue for caution in interpreting positive associations and for better-controlled research to resolve remaining uncertainty.