Acetaminophen (for example, brand name Tylenol) has been widely used for decades to treat pain and fever during pregnancy. Recently, however, concerns have been raised about its safety, particularly regarding whether children exposed to the medication in utero have an increased risk of developing neurodevelopmental disorders, such as autism spectrum disorder (ASD) or attention deficit hyperactivity disorder (ADHD).
Neurodevelopmental disorders such as autism and ADHD are likely the result of multiple genetic and environmental factors, and identifying the specific underlying causes remains a frustrating puzzle for clinicians, researchers, and parents. Therefore, it is an appealing prospect to consider that a clear, causal link could exist between an avoidable risk factor and a specific outcome. Since avoiding acetaminophen during pregnancy in an attempt to reduce the risk of neurodevelopmental disorders in children would limit the options available to treat symptoms like fever and headache, it is important to carefully examine the evidence linking acetaminophen use during pregnancy to neurodevelopmental disorders in children exposed in utero.
Two Studies Fueling Recent Questions About Acetaminophen Safety in Pregnancy
Two studies have been cited to support claims of a potential association between acetaminophen exposure in utero and neurodevelopmental disorders in offspring: the 2019 Nurses’ Health Study II and the 2020 Boston Birth Cohort Study.
The 2019 Nurses’ Health Study II included 8,856 individuals who gave birth between 1993–2005. This study used a questionnaire to assess how often the participants used acetaminophen during the year they gave birth, and whether their child later developed ADHD (the risk of autism was not evaluated). The researchers found that the risk of ADHD was 1.3 times higher in children whose birthing parent regularly used acetaminophen while pregnant. A thirty percent increase in risk could sound alarming, but the context and quality of the data are crucial in determining how trustworthy or clinically important that information may be. In fact, this study had important limitations, several of which are stated by the authors themselves.
First, the reasons for regular use of acetaminophen, such as infection or comorbidities, were not specified. In other words, the study did not control for the potential effects of any medical conditions that might themselves have affected the offspring in utero or might have necessitated the use of other medications. Second, the definition of “regular use” was inconsistent across the data, ranging from at least two times per week to at least one day per week. Third, the survey asked individuals to remember and report behaviors over the past two years, which is a methodology with a high risk of recall bias. Individuals often don’t remember events over a two-year period accurately or completely and using this as a research strategy introduces potential errors that can affect conclusions in critical ways. Finally, ADHD diagnoses were not verified by a clinical diagnosis and were exclusively self-reported by parents.
The second study cited to support an association between acetaminophen and neurodevelopmental disorders, the 2020 Boston Birth Cohort Study, measured acetaminophen metabolites in umbilical cord blood from 996 parent-infant pairs. The researchers found that the presence of acetaminophen metabolites in cord blood was associated with 2.3 times higher odds of autism in offspring. Like the 2019 study, this potentially alarming finding must be considered in terms of the quality of the data. And, like the 2019 study, there are several important limitations.
First, given that the half-life of acetaminophen in adults is less than three hours, the cord plasma measurement may, at most, reflect acetaminophen use just before delivery. This does not provide an accurate picture of the fetus's total in utero exposure to the drug throughout the pregnancy. Second, the study used archived umbilical cord plasma samples that were not collected specifically for the purpose of this study but were collected for some other indication for cord blood testing. Cord blood is typically only collected if there are delivery complications or concerns about fetal wellbeing. This indicates that the sample population may not be representative of the general population, but rather a population with a high risk of complications.
Finally, both cohort studies included fewer than 10,000 individuals, a relatively small sample size for epidemiological studies, which often include millions.
Is there higher-quality data evaluating acetaminophen in pregnancy?
Yes. In fact, the two highest quality studies that investigated this issue found no association between acetaminophen exposure and neurodevelopmental disorders. The first was a 2024 cohort study in Sweden that included nearly 2.5 million offspring from a nationwide birth cohort between 1995 and 2019. The second was a 2025 cohort study in Japan of over 200,000 offspring from a birth cohort between 2005 and 2022, which cleanly replicated the results of the 2024 study. Both studies used data from national registries and rigorous statistical methods to evaluate the association between acetaminophen use during pregnancy with the development of neurodevelopmental disorders in offspring.
The most important distinction between these two large studies and the earlier ones was the inclusion of a sibling-control analysis. This analysis takes into account whether siblings with different exposures to acetaminophen had different rates of neurodevelopmental disorders. Developmental disorders are highly heritable, so the purpose of a sibling-control analysis is to statistically control for the effects of genetics, along with other shared familial factors such as socioeconomic status and early childhood environment. In both studies, when the researchers performed the analysis without accounting for genetic factors, there was a small statistical association between in utero acetaminophen exposure and autism, ADHD, and intellectual disability. But after performing the sibling-controlled analysis, these associations disappeared.
When taken together, the body of evidence does not support a significant or clinically meaningful association between in utero exposure to acetaminophen and the development of autism or ADHD in offspring, and importantly, a causal relationship has not been established by any study.
In all observational studies investigating correlations between risk factors and outcomes, confounding variables may exist that can lead to erroneous conclusions about cause and effect. Indeed, the attention being paid to a potential association between acetaminophen and autism or ADHD highlights the need for careful, rigorous evaluation of scientific evidence by clinicians and experts in evidence assessment and grading. In this case, for example, the conditions that might prompt a pregnant person to use acetaminophen, such as infection or fever, could be confounding as they have been independently associated with increased risk of neurodevelopmental conditions in prior studies. Without performing a randomized controlled trial, it is not possible to establish causality – that is, to establish for certain whether a statistical association between acetaminophen and neurodevelopmental conditions is due to the acetaminophen exposure itself, rather than any of these confounding factors. The problem is, it is essentially impossible to do such a trial due to the ethical implications of knowingly exposing a pregnant person and fetus to a potentially harmful situation such as an untreated fever. Instead, all conclusions must be drawn from observational studies alone, which is why the process of critically appraising the methodology, analysis, and confounding factors in these studies is so vital.
When taken together, the body of evidence does not support a significant or clinically meaningful association between in utero exposure to acetaminophen and the development of autism or ADHD in offspring.
When taken together, the body of evidence does not support a significant or clinically meaningful association between in utero exposure to acetaminophen and the development of autism or ADHD in offspring.
How does the evidence inform clinical practice?
When translating this body of evidence to clinical practice, it is clear that outright avoidance of acetaminophen during pregnancy is unnecessary at best and could even be harmful. Not only may untreated pain or fever itself be harmful, but if acetaminophen is not used, patients may turn to other medications that have known harmful effects. The American College of Obstetricians and Gynecologists (ACOG) reaffirmed in September that acetaminophen remains the safest analgesic option to use during pregnancy. Like any medication use in pregnancy, the best practice is still to use acetaminophen at the lowest possible dose for the shortest duration necessary. The data do not suggest that acetaminophen is harmful for pregnant people or their children. Clinicians should feel empowered to lean on the evidence to alleviate concerns or confusion that their patients might have regarding this issue.
