Lead
A new umbrella review published in BMJ concludes that current evidence does not clearly establish a causal link between acetaminophen (paracetamol) use in pregnancy and later diagnoses of autism spectrum disorder or attention-deficit/hyperactivity disorder in children. The authors synthesized nine systematic reviews that together summarized 40 primary studies and found consistent methodological weaknesses across reviews and studies. Several reviews reported possible associations, but most cautioned against concluding causation. The review’s authors and major medical societies warn that incomplete evidence should not prompt pregnant people to avoid treatment of high fever or severe pain without clinical advice.
Key Takeaways
- Nine systematic reviews were examined; those reviews covered 40 individual studies of prenatal acetaminophen exposure and child neurodevelopment.
- Seven of the nine reviews explicitly warned that their findings did not prove a causal relationship between maternal acetaminophen use and autism, ADHD, or both.
- Reviewers identified common flaws: incomplete literature searches, missing exclusion lists, inconsistent study-design classification and nonstandard bias-assessment tools.
- Only one of the nine reviews included sibling-controlled studies; those sibling analyses removed the apparent association when accounting for shared familial factors.
- The review team comprised researchers from the U.K., Spain and Australia and rated the nine reviews as low or critically low confidence for summarizing the available evidence reliably.
- Public statements that acetaminophen use in pregnancy is definitively linked to autism or ADHD are not supported by the overall quality of the evidence, according to the BMJ umbrella review.
- Clinicians caution that untreated fever or severe pain in pregnancy can carry risks for mother and fetus; ibuprofen and some alternatives are not safe in pregnancy, complicating public-health advice.
Background
Acetaminophen (paracetamol) is one of the most commonly recommended analgesics for use in pregnancy because, compared with many alternatives, it has been considered relatively safe in standard dosing. In August 2025 a large meta-analysis involving researchers from Mt. Sinai, Harvard T.H. Chan, UCLA and UMass Lowell reported prenatal acetaminophen exposure may be associated with higher rates of neurodevelopmental diagnoses; that analysis prompted renewed scrutiny and debate.
The topic moved into high-profile public discussion in September when a U.S. political leader publicly warned pregnant people about acetaminophen and recommended avoidance, a statement that several major medical groups, including the American College of Obstetricians and Gynecologists, criticized as lacking support from the full body of scientific evidence. The BMJ umbrella review was undertaken in part to assess the overall quality and consistency of existing reviews and to provide clinicians and patients a clearer synthesis.
Main Event
The BMJ paper pooled and critically evaluated nine published systematic reviews that themselves summarized 40 primary observational studies of prenatal acetaminophen and later child neurodevelopmental outcomes. The umbrella review team assessed each review’s search strategy, inclusion/exclusion transparency, assessment of study designs and risk-of-bias tools, finding widespread methodological limitations across reviews.
While the nine reviews commonly reported a potential association between maternal acetaminophen use and adverse neurodevelopmental outcomes, the BMJ authors found that seven reviews explicitly warned against inferring causation. The umbrella review highlighted that many primary studies did not adequately adjust for confounding factors such as genetic predispositions, maternal health, indication for medication (fever or pain), and environmental exposures.
Crucially, only one of the nine reviews included sibling-controlled studies—analyses comparing siblings with differing prenatal exposures—to help separate effects of the medication from familial or genetic confounders. In those sibling-controlled analyses, the previously observed associations attenuated or disappeared, suggesting shared familial or environmental factors may explain initial signals.
The BMJ authors rated the nine reviews overall as low or critically low confidence, meaning the reviews could not reliably be taken as comprehensive or definitive summaries of the evidence. They conclude that the existing literature does not offer a clear causal pathway linking prenatal acetaminophen with autism or ADHD.
Analysis & Implications
Interpretation requires a careful distinction between correlation and causation. Observational studies can identify statistical associations, but residual confounding—unmeasured or imperfectly measured factors such as genetics, maternal indications for use, socioeconomic status, or environmental exposures—can create false impressions of a drug effect. The umbrella review emphasizes that many studies failed to control adequately for these influences.
Sibling-control designs are one robust approach to address shared familial confounding because they compare siblings who are differently exposed in utero. The fact that sibling-adjusted analyses removed associations in the one review that included them increases the likelihood that earlier observed links reflected familial factors rather than direct drug toxicity.
Clinically, the review cautions against blanket messaging to stop acetaminophen use in pregnancy. High fevers and severe pain can have harms for both mother and fetus; if patients avoid acetaminophen without medical guidance, they may risk undertreatment or substitute potentially unsafe alternatives such as ibuprofen, which is contraindicated at certain stages of pregnancy.
Policy implications include the need for higher-quality data: prospective studies with detailed indication data, standardized exposure measurement, genetic and environmental covariates, and replication of sibling-control designs. The review also highlights an ethical and scientific gap—pregnant people are often excluded from drug trials, leaving clinicians to rely on imperfect observational datasets for safety guidance.
Comparison & Data
| Metric | Count |
|---|---|
| Systematic reviews assessed | 9 |
| Primary studies across reviews | 40 |
| Reviews advising caution on causation | 7 of 9 |
| Reviews including sibling analyses | 1 of 9 |
| Overall confidence rating (umbrella review) | Low or critically low |
The table summarizes the umbrella review’s key counts. These headline figures illustrate why the authors judged the evidence base insufficient for a causal claim: relatively few studies have robust designs that address shared familial confounding, and many reviews relied on overlapping primary datasets.
Reactions & Quotes
Stakeholders responded quickly after the BMJ paper’s publication, reflecting differing emphases on public caution and scientific nuance.
“We assessed the breadth and quality of existing reviews to help clinicians discuss risks and benefits with pregnant women and families.”
Dr. Shakila Thangaratinam, University of Liverpool (co-author)
Dr. Thangaratinam stressed that the goal was to evaluate methodological quality and to avoid prompting pregnant people to abruptly stop treatment for fever or severe pain without clinical advice.
“Public statements that imply a definitive causal link are highly concerning and not backed by the full body of evidence.”
American College of Obstetricians and Gynecologists (ACOG) — professional society statement
ACOG and other major medical groups have urged that guidance remain evidence-based and that clinicians counsel patients on risks of untreated maternal fever and pain.
“Sibling-controlled analyses suggest family and genetic factors may account for the associations reported in many studies.”
Independent epidemiologist (expert analysis)
Experts reiterated that designs addressing familial confounding shift interpretation away from direct drug causation toward shared risk explanations.
Unconfirmed
- The extent to which public statements in September directly changed medication behavior among pregnant people is not established in the reviewed evidence and remains unquantified.
- Links reported in some primary studies between acetaminophen and specific neurodevelopmental diagnoses may be influenced by unmeasured indications for use (for example, maternal infection or fever) rather than the medication itself; this remains unresolved in many datasets.
- Whether particular dosing patterns, timing during pregnancy or cumulative exposure pose measurable risks has not been established and requires better prospective data.
Bottom Line
The BMJ umbrella review finds that while multiple systematic reviews report possible associations between prenatal acetaminophen use and autism or ADHD, methodological weaknesses and lack of causal evidence mean a clear causal link is not established. Sibling-controlled analyses—available in only a small subset of the literature—attenuate the apparent association, suggesting shared familial or genetic factors may explain earlier signals.
For clinicians and pregnant people, the practical takeaway is to balance risks: do not rely on headlines to stop treating high fever or severe pain without professional guidance, and avoid substituting medications that are not recommended in pregnancy. The field needs better-designed prospective studies, more consistent exposure measurement, and inclusion of pregnant people in clinical research to guide safer therapeutic options.
Sources
- ABC News — news report summarizing the BMJ review and public reaction (media)
- BMJ (British Medical Journal) — peer-reviewed journal; source of the umbrella review (academic journal)
- American College of Obstetricians and Gynecologists (ACOG) — professional society statement responding to public claims (medical society)
- Mt. Sinai Health System / affiliated research — institutions involved in the August 2025 meta-analysis referenced in coverage (academic research institutions)