When Public Claims Outpace Causation
Public debate over alleged links between prenatal acetaminophen exposure and autism spectrum disorder has resurfaced with renewed intensity. Once again, scientific research has been drawn into political and legal argument, often without adequate attention to evidentiary standards or methodological limits. For expert witnesses, this is familiar territory.
A recent systematic review published in The Lancet Obstetrics, Gynecology & Women’s Health offers an opportunity to reset the discussion. The authors examined 43 observational studies assessing associations between acetaminophen (paracetamol) use during pregnancy and later neurodevelopmental outcomes, including autism spectrum disorder, attention-deficit/hyperactivity disorder, and intellectual disability. Their conclusion was straightforward: when the most rigorous study designs are given appropriate weight, the evidence does not support a causal relationship.
That distinction matters. In both litigation and regulatory settings, the difference between association and causation is not academic. It is determinative.
Study Design and the Hierarchy of Evidence
The Lancet review did not simply tally studies for or against an association. Instead, it applied a hierarchy of methodological reliability. Studies using sibling-comparison designs were given particular emphasis, as these control for shared genetic, environmental, and familial factors that often confound epidemiological research.
From an expert witness perspective, this approach aligns closely with judicial expectations. Courts do not treat all studies as equal. Designs that reduce bias, limit confounding, and permit meaningful inference are consistently afforded greater probative value.
Dr. Asma Khalil, the review’s lead author and a fetal medicine specialist at St. George’s Hospital in London, summarized the findings plainly: acetaminophen remains safe to use during pregnancy and continues to be recommended as first-line treatment for pain and fever. That conclusion was not based on absence of data, but on careful evaluation of its quality.
In practice, when sibling-comparison studies were prioritized, the purported association between acetaminophen exposure and autism was not observed.
Confounding, Bias, and the Limits of Observational Research
Many studies reporting small statistical associations fail to adequately address confounding variables. Pregnant individuals do not take acetaminophen at random. It is commonly used to treat fever, infection, and inflammatory conditions—each of which may independently increase neurodevelopmental risk.
This is not a trivial issue. If the underlying condition prompting medication use is itself a risk factor, failure to control for it can produce misleading results. In medico-legal terms, this undermines any opinion offered on general causation.
Recall bias further complicates interpretation. Several studies rely on retrospective self-reporting by mothers, often years after pregnancy and after a diagnosis has been made. Courts have long recognized the limitations of such evidence. Memory is imperfect, and retrospective exposure assessment is particularly vulnerable to distortion.
When these sources of bias are addressed using more robust methodologies, the claimed association weakens or disappears entirely.
Selective Reliance and Disclosure in Litigation-Adjacent Research
Despite the broader body of evidence, public attention has focused on a 2024 review published in BMC Environmental Health. That review analyzed 46 prior studies and concluded that they supported an association between prenatal acetaminophen exposure and increased incidence of neurodevelopmental disorders.
Several methodological concerns have been raised about that analysis, including heterogeneity of outcomes, limited differentiation between diagnoses, and inclusion of smaller studies prone to bias. Only a subset of the cited research addressed autism specifically.
Of particular relevance to expert witnesses is the disclosure made by the review’s senior author, Dr. Andrea Baccarelli, Dean of the Harvard T.H. Chan School of Public Health. In the paper, he noted that he has served as an expert witness for plaintiffs in litigation involving alleged links between acetaminophen use during pregnancy and neurodevelopmental harm.
Disclosure does not invalidate research. However, in courtroom settings, such disclosures inevitably prompt heightened scrutiny. Judges and juries are entitled to consider whether conclusions align with methodological rigor or litigation posture.
That scrutiny is especially warranted where conclusions are framed as supporting causation, despite explicit acknowledgments that association alone does not establish cause and effect.
Genetics and Established Risk Factors
One point on which experts broadly agree is that genetics represent the most significant risk factor for autism spectrum disorder. Additional factors—including advanced paternal age, preterm birth, and maternal health complications—are also well documented.
High-quality epidemiological studies routinely account for these variables. Lower-quality studies often do not. From an expert testimony standpoint, failure to control for known risk factors materially weakens any causal inference.
It is also worth noting that no plausible biological mechanism has been conclusively established by which therapeutic acetaminophen use, at recommended doses, would independently cause autism. While biological plausibility is not strictly required, its absence further limits the strength of causation claims.
Public Health Consequences of Overcorrection
A commentary published alongside the Lancet review, authored by researchers from the London School of Hygiene & Tropical Medicine, Children’s Hospital Colorado, and other institutions, raised a separate concern. Discouraging acetaminophen use during pregnancy, they cautioned, may lead to inadequate treatment of pain, fever, or infection.
That risk is not hypothetical. Untreated maternal fever and infection are associated with well-established adverse fetal outcomes, including neurodevelopmental harm. From a risk-benefit perspective, discouraging a widely used medication without compelling causal evidence may itself cause preventable harm.
For expert witnesses, this underscores a broader responsibility: opinions offered in litigation do not exist in isolation. They influence public understanding, clinical practice, and policy decisions.
Conclusion: Evidence, Expertise, and Restraint
The debate surrounding acetaminophen and autism illustrates the importance of evidentiary discipline. Observational studies can generate hypotheses, but they cannot substitute for rigorous causal analysis. Courts, regulators, and the public deserve conclusions that reflect the full weight of evidence, not its most provocative elements.
When the most methodologically sound studies are prioritized—particularly those accounting for shared genetics and environment—the alleged association between prenatal acetaminophen exposure and autism is not supported.
As Dr. Khalil observed, when appropriate analytical frameworks are applied, “the association is not seen.” That finding reflects both sound science and the level of proof required when science is asked to serve society, the courts, and the individuals whose lives may be affected by its interpretation.
https://www.acog.org/news/news-releases/2025/09/acog-affirms-safety-benefits-acetaminophen-pregnancy
https://witnessdirectory.com/signup.php
