The hullaballoo over the acetaminophen (known in South Africa as paracetamol) and autism issue, and the warnings issued by the Trump administration on its use for pregnant women, have created some confusion and conflict, writes Stephen Hoption Cann in Medscape – who cautions that doctors should be transparent about the current state of evidence.
“While research has yet to reach a definitive conclusion, it would be premature to offer blanket reassurance about the safety of this drug during pregnancy,” he advises.
He writes:
On April Fool’s Day in 1961, thalidomide was approved for sale in Canada as a treatment for morning sickness in pregnant women. Within less than a year, it was withdrawn from the market as alarming evidence of severe birth defects began to emerge. Among the most devastating effects was phocomelia, a condition in which babies were born with absent or severely shortened, often flipper-like limbs.
In addition to limb malformations, other serious abnormalities were reported, including facial and cranial deformities, as well as defects in internal organs.
South of the border, a battle over thalidomide’s approval unfolded, but it was never authorised in the United States, thanks to the steadfast opposition of US Food and Drug Administration (FDA) reviewer Frances Kelsey, MD, PhD.
Her ongoing concerns about the drug’s safety were ultimately validated when reports from Europe linked thalidomide use during pregnancy to a surge in births of severely deformed infants. This public health crisis prompted the passage of the Kefauver-Harris Amendment, which significantly strengthened drug regulation in the US by requiring manufacturers to provide robust evidence of safety, including results from animal testing, before receiving marketing approval.
Tragically, although thalidomide was never officially approved in the US, it was nonetheless distributed widely through loosely regulated clinical trials, exposing thousands of patients, including pregnant women, to its devastating effects.
This emphasis on birth defects, however, meant that drug safety assessments often overlooked the potential for long-term health effects – a concern now gaining increasing attention for many medications long assumed to be safe during pregnancy.
Acetaminophen, which is widely considered to entail low risk during pregnancy, is one such drug; however, a growing number of studies have linked its use to adverse neurodevelopmental outcomes, including autism.
Once rarely used during pregnancy, the drug saw a dramatic rise in popularity after studies linked aspirin use in children to Reye syndrome, a condition marked by cerebral oedema and fatty liver degeneration. In response, the FDA issued an advisory in 1982 warning against aspirin use in children with flu or chickenpox.
By 1986, it mandated that all aspirin products carry a warning label about the risk for Reye syndrome.
As aspirin use in children rapidly declined in the early 1980s, acetaminophen quickly became its replacement. Around the same time, autism rates began to rise sharply. Intrigued by this parallel, Becker and Schultz examined the correlation between the two trends during that decade.
While both acetaminophen use and autism showed a consistent upward trajectory, there were two notable exceptions: brief declines in autism rates after the 1982 and 1986 Tylenol tampering incidents, in which cyanide-laced capsules were placed on pharmacy shelves. Although no perpetrator or motive was ever confirmed, public fear led to temporary drops in acetaminophen use during each episode.
In 1970, the prevalence of autism was estimated at four in 10 000, but by 2022, US Centres for Disease Control and Prevention data reported it at roughly one in 31. Of course, times have changed since the 1970s: autism is recognised more broadly as autism spectrum disorder (ASD), diagnostic criteria are more precise, and increased awareness has contributed to this rise.
Still, many experts believe the growing case numbers reflect a genuine increase in prevalence, not just for ASD but also for other neurodevelopmental disorders. Since Becker and Schultz’s initial study, numerous others have explored the potential link between acetaminophen use during pregnancy and neurodevelopmental outcomes.
Most of these findings remained relatively obscure until the 2021 publication of a consensus statement in Nature Reviews Endocrinology by 91 scientists, clinicians, and public health professionals from around the world. The authors warned that “increasing experimental and epidemiological research suggests that prenatal exposure to APAP (acetaminophen) might alter foetal development, which could increase the risks of some neurodevelopmental, reproductive, and urogenital disorders”.
They further advised that pregnant women should be cautioned early in pregnancy to avoid acetaminophen unless medically indicated, consult a physician or pharmacist if uncertain about its necessity or before long-term use, and minimise exposure by using the lowest effective dose for the shortest possible duration.
In response to the 2021 consensus statement, the Society of Obstetricians and Gynaecologists of Canada (SOGC) issued a rebuttal, stating that “the evidence for causality for this claim is weak and has many fundamental flaws”. To support their position, they cited a 2015 US FDA Drug Safety Communication on the risks of pain medicine (acetaminophen, nonsteroidal anti-inflammatory drugs, and opioids) use during pregnancy.
The FDA concluded that “all of the studies we reviewed (had) potential limitations in their designs” and that conflicting results prevented reliable conclusions. As a result, the FDA did not alter its recommendations for these drugs during pregnancy.
Regarding acetaminophen specifically, the FDA referenced a 2014 cohort study by Liew and colleagues, which found that prenatal acetaminophen use was associated with increased risks for hyperkinetic disorders and attention-deficit/hyperactivity disorder (ADHD)-like behaviours in children; however, the study’s authors emphasied that while the findings were of public health relevance, further research was needed to confirm causality.
The SOGC statement also cited a 2017 review by the Society for Maternal-Foetal Medicine Publications Committee, which concluded that “the weight of evidence is inconclusive regarding a possible causal relationship between acetaminophen use and neurobehavioral disorders in the offspring”.
To support this conclusion, they referenced two comprehensive reviews. The first, by Hoover and colleagues (2015), found a mild correlation between prenatal acetaminophen use and ADHD symptoms in children but concluded that the available data were insufficient to establish a causal link.
The second paper, by Andrade (2016), concluded that the studies reviewed suggest that prenatal exposure to acetaminophen increases the risk for ADHD and this association may be dose-dependent, though the author emphasised the need for further research to confirm these findings.
Researchers continued to examine this potential association, and a 2024 population-based study by Ahlqvist and colleagues gained prominence. Their analysis included comparisons with and without sibling controls. In the general population analysis, they observed marginally increased risks for autism and ADHD. But these associations were no longer evident when matched sibling pairs were used.
Similarly, while some of their models suggested a dose-response relationship between acetaminophen use and outcomes such as autism, ADHD, or intellectual disability, these findings also disappeared in the sibling-controlled analysis. The authors proposed that unmeasured confounding factors might explain the discrepancy, though the nature of such confounders remains unknown.
Finally, this summer, a comprehensive systematic review was published by Prada and colleagues synthesising research on acetaminophen use and neurodevelopmental disorders (NDDs) up to 25 February 2025.
Their review incorporated evidence from human studies, animal models, and mechanistic research. After exclusions, 46 studies were retained for further analysis. Overall, 27 studies observed significant links to NDDs, nine found no significant link, and four suggested potential protective effects.
Notably, the authors observed that studies with higher methodological quality were more likely to report positive associations. They concluded that the findings “demonstrated evidence consistent with an association between exposure to acetaminophen during pregnancy and offspring with NDDs”.
They emphasised the need for further research to confirm these associations and clarify underlying mechanisms, however.
Prada and colleagues were critical of sibling-controlled studies, including the one conducted by Ahlqvist and colleagues, because of the inherent limitations in sample size and statistical power. In sibling comparison designs, only the pairs where one sibling was exposed and had the outcome, while the other was unexposed and did not develop the outcome, contribute meaningfully to the analysis.
If both siblings were exposed, or both had the same outcome, this does not help answer the research question and greatly reduces the number of usable cases, thereby diminishing statistical power. While sibling-controlled designs offer the advantage of accounting for shared familial and genetic factors, they may be more susceptible to random exposure misclassification.
This is especially problematic when maternal behaviours are assumed to vary between pregnancies, despite evidence suggesting that they may remain consistent. For instance, Ahlqvist and colleagues reported an unusually low rate of acetaminophen use during pregnancy (7.49%), raising concerns about potential exposure misclassification and its impact on the validity of the findings.
The review by Prada and colleagues prompted a new response from the SOGC. In its statement, the SOGC indicated that it had revisited the evidence and maintained that “evidence for causality is weak and that the results of the studies claiming a causal link should not alter current clinical practice”.
To support their position, they cited two recent publications: the sibling-controlled study by Ahlqvist and colleagues (referenced earlier), and a review by Damkier and colleagues published online in December 2024.
Based on their analysis, the authors concluded that “in utero exposure to acetaminophen is unlikely to confer a clinically important increased risk of childhood ADHD or ASD”.
Notably, in contrast to the assessments by Prada and colleagues, their review rated population-based studies as generally methodologically poor, while sibling-matched designs were rated more favourably, highlighting the ongoing debate over study design and data reliability in this area of research.
The broader scientific disagreement surrounding prenatal acetaminophen exposure and neurodevelopmental outcomes has led to the emergence of two distinct interpretive camps, which unhelpfully have come to opposite conclusions from the available evidence.
The US Department of Health and Human Services has now weighed in on this debate, stating that “evidence does not definitively establish causality, but the consistent associations raise concern”.
In response, the FDA is initiating a label change “to make parents and doctors aware of a considerable body of evidence about potential risks associated with acetaminophen”.
The FDA adds: “Even with this body of evidence, the choice still belongs with parents. The precautionary principle may lead many to avoid using acetaminophen during pregnancy, especially since most low-grade fevers don’t require treatment.”
For better or worse, a real-world experiment is now under way. Many women contemplating pregnancy, who may not have been previously aware of this ongoing debate, will now be informed. This heightened awareness is likely to lead to a decline in acetaminophen use during pregnancy.
Whether this shift will result in a corresponding change in the incidence of NDDs remains to be seen. In any case, more comprehensive studies are currently under way or being planned, and they should offer more definitive answers to this important question.
For clinicians, this issue can no longer be avoided in patient discussions. Healthcare providers should be transparent about the current state of evidence.
Research has yet to reach a definitive conclusion, and patients deserve to know that uncertainty remains, and that ongoing studies may shift our understanding.
Presenting the available information honestly allows patients to make informed choices, even in the face of incomplete data.
Stephen A. Hoption Cann, PhD, is a clinical professor and epidemiologist in the School of Population and Public Health at the University of British Columbia in Vancouver. His research focuses on adverse drug effects and infectious disease management.
Medscape article – Acetaminophen and Autism: What Do the Data Suggest? (Open access)
See more from MedicalBrief archives:
SA expert groups dismiss Trump's paracetamol link to autism
Kennedy to ‘reveal causes of autism’ this month
Paracetamol link to autism rears again