Since the Department of Health and Human Services (HHS) warned against acetaminophen use in pregnancy because of autism risks in September 2025, several contradicting papers have come out in response to the announcement.
A study published on Jan. 16 in The Lancet Obstetrics, Gynaecology, & Women’s Health is the latest addition to the body of research pushing back.
A pattern emerges when looking at the studies that disapprove of the acetaminophen–autism link; most of them support or have heavily cited the findings of a 2024 Swedish study.
Acetaminophen is the active ingredient in Tylenol.
The Lancet Study
William Parker, who worked at Duke University for more than 27 years and is now a visiting scholar at the University of North Carolina and CEO of WPLab Inc., said the Lancet study is one of several studies pushing back against the U.S. government’s warning against acetaminophen.
Some of these studies are newly conducted human trials, but many are reviews, which compile previous studies’ findings.
The Lancet study is a review that analyzed 43 papers using high-quality and rigorous research methods, according to a statement.
The researchers looked at sibling-comparison studies, designed to compare only siblings and thus eliminate potential confounding factors such as genetics and environment.
Notably, the review gave a 97 percent weighting to a 2024 study that analyzed 2.5 million children conducted by researchers from Sweden’s Karolinska Institutet and Drexel University.
Dr. Asma Khalil, professor of obstetrics and maternal and fetal medicine at City St George’s–University of London and consultant obstetrician, who led the study, clarified that because the Swedish study examines a larger population than other studies, this naturally affects its weighting.
She said the Lancet study’s conclusions were not based on a single study; her team also reviewed other studies and considered their findings when writing the report.
Parker said, “People perceive that [the Swedish study] has the best data, and it has the most comprehensive data.”
Apart from the Lancet study, the Swedish study has been cited by several reviews published after the HHS announcement in September that challenge the agency’s warning against prenatal acetaminophen use.
These include a BMJ umbrella review, a review in the Journal of the American Academy of Child & Adolescent Psychiatry, and a review in the Journal of Clinical Psychiatry. All conclude that prenatal acetaminophen use is not associated with increased risk.
However, the findings of the Swedish study have come under debate in recent years.
The Debate on the Swedish Study
The Swedish study is a nationwide sibling-comparison study that included 2.48 million Swedish children born between 1995 and 2019.
The study analyzed prescriptions of acetaminophen and found that 7 percent of pregnant mothers used acetaminophen.
The raw data showed a link between acetaminophen use and autism, attention-deficit/hyperactivity disorder, and intellectual disability. However, this link weakened and then disappeared when researchers adjusted for the sex of the offspring, genetics, and maternal health.
In the HHS’s September announcement against acetaminophen use, the agency cited a review led by Dr. Andrea Baccarelli, dean of the Harvard T.H. Chan School of Public Health, which discussed potential flaws in the Swedish study.
In that Harvard study, Baccarelli and his authors found the Swedish study to be highly biased. One critique was that the study relied on acetaminophen prescription data, which could underestimate actual acetaminophen use.
Because acetaminophen is often purchased over the counter, some pregnancies may have been classified as not taking acetaminophen even though the mother used it.
His team also pointed to other Swedish studies conducted at the same time, in which blood tests and self-reports showed that up to 63 percent of mothers took acetaminophen during pregnancy.
However, the Lancet study found the Harvard study to be biased.
Khalil, in an email to The Epoch Times, said that the issue that the Harvard study raised was considered in her paper.
She added that prescription-based exposure and over-the-counter exposure are different.
For example, prescription-based acetaminophen is often prescribed at a higher dose than over-the-counter acetaminophen, so it can have a different effect on the pregnancy.
Another, less acknowledged critique that Parker raised is that the Swedish paper excessively excluded confounding factors.
Parker said that science often uses a reductionist lens, isolating factors that in reality work together.
“All the things that they’re adjusting for, including genetics, [the Swedish researchers] see it as confounding factors, but all of those things interact with acetaminophen,” he said.
Genes affect how children metabolize drugs, which can predispose certain children to acetaminophen toxicity.
Researchers also adjusted for maternal health, such as whether the mother smoked, whether she was obese, or whether she had certain metabolic health problems.
Parker said maternal health also affects the effects of acetaminophen. When acetaminophen is metabolized, it releases reactive oxygen species that cause stress on the body. Mothers with poor health would have a harder time metabolizing the drug, which can make the fetus more vulnerable to acetaminophen toxicity.
“There’s a lot of other studies that show very, very similar things, where the more you adjust for factors, the lower the risk goes, and it’s just because those factors are super important,” Parker said. “It’s essentially the same as proving that it’s OK for babies to play with matches, but only if there’s nothing flammable in the area and they’re wearing fireproof clothing.”
Sibling Studies Conundrum
Parker’s critique of the Swedish study pertains specifically to its design.
As a sibling-comparison study, its design requires researchers to control for shared familial and genetic factors and shared maternal characteristics, the very factors that Parker said should not be controlled.
However, precisely because sibling-comparison studies require these adjustments, the broader academic community recognizes sibling studies to be the most rigorous.
“While some argue that these factors may interact rather than confound, the approach is widely accepted as one of the strongest methods for addressing confounding,” Khalil told The Epoch Times.
Beyond the Swedish study, her team also evaluated other sibling-comparison studies, which showed a similar pattern: no link between prenatal acetaminophen use and autism.
Sibling-comparison studies have generally found no evidence of harm from acetaminophen. By contrast, population-level observational studies, which cannot be as rigorously controlled, including the Nurses’ Health Study II and the Boston Birth Cohort, have sometimes reported associations.
Other experts suggest that the design of sibling studies may wash out the true effects of drugs.
In an email to The Epoch Times, Andrew Nixon, an HHS spokesperson, pointed to an X post in which National Institutes of Health Director Dr. Jay Bhattacharya outlined reasons why the reliability of sibling-comparison studies is overhyped.
“Sibling studies focus on siblings who differ in treatment (e.g., one took a drug, the other didn’t),” he wrote. This excludes families with only one child and families in which siblings all took the same drug, which can lead to weaker findings, as the sample size would be small.
The second issue is that families in which siblings differ in treatment are not random, which can introduce new biases, he said.
For example, it could be that for one pregnancy the mother got sick and took acetaminophen but that she was healthy for her next pregnancy and did not take it. These short-term differences in maternal health are not controlled for in sibling studies and can introduce bias.
“Sometimes the very factors controlled for (genes, environment) are part of how the treatment causes the outcome,” Bhattacharya said. “Sibling studies can ‘wash out’ real effects, making a true link look like nothing.”
The Conundrum for Mothers
At the end of the day, acetaminophen is the first-choice painkiller for women during pregnancy.
Avoiding acetaminophen even when experiencing significant pain or fever can expose both mother and baby to known risks, particularly untreated maternal fever, the Lancet researchers said.
The HHS said that when used short-term and at appropriate dosages, acetaminophen remains widely regarded as safe.
There is a gold-standard treatment for acetaminophen overdose. N-acetylcysteine (NAC) helps replenish antioxidants lost from metabolizing acetaminophen.
NAC is generally given after acetaminophen overdose; there are limited studies on taking NAC prophylactically.
Unresolved questions remain regarding chronic or late-pregnancy acetaminophen use, according to the HHS.