A box of Tylenol photographed in 2015.

Can In-Utero Exposure to Acetaminophen Cause Autism and ADHD?

More than 100 families of children with autism and attention deficit hyperactivity disorder are suing companies that market acetaminophen, the pain reliever in Tylenol and an array of other medications. Tylenol-maker Johnson & Johnson, as well as major retailers that use acetaminophen in their store-brand products, knew about research linking prenatal use of acetaminophen to neurodevelopmental disorders in children, the families claim, and should have included warnings on product labels.

The court filings reveal mothers wracked with guilt, convinced that by taking an over-the-counter pain reliever, they caused their child’s disability. In case after case, these women say that if they had thought acetaminophen could possibly harm their baby, they would have minimized their use of the drug — or not taken it at all.

It’s hardly an open-and-shut case. Most of what is known about acetaminophen and pregnancy comes from a type of study that sifts through data looking for correlations between prenatal exposures and developmental disorders. Scientists have been fighting amongst themselves over how much weight to give these studies, which were not designed to prove that a given factor — in this case, acetaminophen — caused ADHD or autism.

The debate reached a boiling point in 2021, when a group of international scientists declared that the current research, limited as it is, warrants stronger warnings about the use of acetaminophen during pregnancy. In a consensus statement published in Nature Reviews Endocrinology, the scientists called for “precautionary action” through focused research and increased awareness of the drug’s potential risks. Ninety-one scientists, clinicians, and public health professionals from around the world signed on.

That statement was the “galvanizing incident” for the lawsuits, said Ashley Keller, a founding partner at the legal firm Keller Postman LLC and one of the lawsuits’ lead attorneys.

But there’s a hitch: The consensus statement does not, in fact, reflect the views of many experts or of any major medical organization. The same Nature journal published three rebuttals signed by numerous professional groups as well as individual researchers and clinicians. These critics wrote that the consensus statement used flawed data to exaggerate potential harms of acetaminophen and downplayed the drug’s essential role for treating fever and pain.

Acetaminophen is one of the most common drugs in the world, and in the United States, up to 65 percent of all expectant mothers use it.

Johnson & Johnson has seized on those criticisms in its defense. The consensus statement “is an outlier opinion of a small group whose position has been rejected by their own medical organizations and every regulatory body to address the issue,” company spokesperson Melissa Witt told Undark in an email. Giving credence to theories not based in sound science, she said, could harm millions of pregnant women.

This debate over how to interpret the acetaminophen science lies at the heart of the lawsuits, and the answer has profound implications, both for individuals and for public health. Acetaminophen is one of the most common drugs in the world, and in the United States, up to 65 percent of all expectant mothers use it. The cases have been consolidated in the Southern District of New York. If the litigation proceeds to trial, attorneys expect tens to hundreds of thousands of families to join in.

It’s a shame that questions linger about the safety of a drug that’s been around for 70 years, said Christina Chambers, a professor of pediatrics at the University of California, San Diego and lead investigator for a series of studies on exposures during pregnancy for the Organization of Teratology Information Specialists, a professional society that provides advice on using medications during pregnancy and breastfeeding.

In an interview with Undark, Chambers expressed doubts about the consensus statement, saying that if acetaminophen has any effect on fetal development, that effect is likely to be modest. Still, she added, “What this accumulated data calls for is to do better study.”

Acetaminophen was discovered in 1878, according to an FDA history of groundbreaking medications. But it wasn’t until the early 1950s that researchers demonstrated that the compound worked as well as the two popular pain relievers of the time — aspirin and acetanilide — with fewer side effects. In 1955, drugmaker McNeil gained FDA approval and launched Tylenol Elixir for Children, advertising it “for little hotheads.” Four years later Johnson & Johnson acquired McNeil, and, in 1975, began aggressively marketing an “extra strength” 500-milligram version of the drug to adults. By the early 1980s, Tylenol was the top-selling pain reliever in the U.S.

Acetaminophen is now used in more than 600 over-the-counter and prescription medications, including combination products for sleep, cough, cold, and allergy.

Acetanilide was one of two popular pain relievers at the time acetaminophen was discovered in 1878. Like many older drugs, acetaminophen didn’t undergo as thorough of safety testing as today’s drugs.

Visual: FDA

Today, new medications typically undergo toxicity testing in animals before researchers study their safety and effectiveness in humans, but like many older drugs, acetaminophen didn’t undergo as thorough of a process. “Back in ’55 we weren’t doing preclinical testing for reproductive safety,” said Chambers.

Scientists do know that, in most cases, acetaminophen is the safest way for pregnant women to relieve fever and pain. “Not treating a fever — especially a high fever — can have consequences,” continued Chambers. Strong data from lab animal and human studies have associated a high fever at certain points in pregnancy with an increased risk of birth defects and other fetal abnormalities. And, while it’s an understudied area, she said, chronic pain is associated with depressive symptoms, insomnia, and other harms to the mother and could adversely affect her pregnancy.

Some common over-the-counter drugs treat both fever and pain, such as nonsteroidal anti-inflammatory drugs like aspirin, ibuprofen (Advil), and naproxen (Aleve), but the FDA warns against using any of these after 20 weeks of pregnancy because these drugs can damage the unborn babies’ kidneys.

Meanwhile, other painkillers have their own drawbacks. Babies exposed to opioids during the first trimester of pregnancy are at increased risk for birth defects of the brain, spine, and spinal cord. In addition, studies show that regular use of opioids during pregnancy increases the risk of poor fetal growth, preterm delivery, stillbirth, and neonatal opioid withdrawal syndrome.

Medicinal cannabis may address pain, and some women might think of it as a natural and, therefore, safe alternative. But there’s very little data available on short- and long-term outcomes for mothers and babies exposed to the potent products available today, said Chambers. Recent studies suggest that cannabis increases the risk of preterm birth and smaller babies, but it’s hard to tease out the effects from other factors. She described the drug’s legalization and increased societal acceptance as “a huge experiment happening now.”

“So, what are you left with?” she asked.

In 1975, Johnson & Johnson began aggressively marketing extra-strength Tylenol to adults. In this 1983 commercial, the product is touted as the most potent pain reliever that can be bought without a prescription.

For pregnant people experiencing fever or pain, acetaminophen is widely viewed as the best option. But can it also harm the fetus?

To begin answering this question, researchers have analyzed preexisting datasets of health information, looking for associations between acetaminophen use during pregnancy and neurodevelopmental problems in children. Such research is referred to as observational, and while it can be useful, this approach can’t typically prove causality.

Autism rates have climbed steadily since the Diagnostic and Statistical Manual of Mental Disorders, or DSM, first established it as a distinct disorder in 1980. In 2000, 1 in 150 children were diagnosed with autism by the age of eight according to the Centers for Disease Control and Prevention; by 2020, the number had risen to 1 in 36. ADHD rates increased as well, though not as sharply. In 2019, 6 million children between the ages of 3 and 17 (9.8 percent) had received a diagnosis of ADHD, compared to 4.4 million children in 2003, according to CDC data from a national survey of parents.

If acetaminophen has any effect on fetal development, that effect is likely to be modest, said Chambers. Still, she added, “What this accumulated data calls for is to do better study.”

Many experts attribute the bulk of that increase to greater awareness and broadening definitions of the disorders. Factors such as improved survival for premature infants and a trend toward starting families later may also play a role, as both prematurity and older parents are associated with increased risk for neurodevelopmental disorders.

In the early 2010s researchers additionally became interested in whether acetaminophen, so commonly used by pregnant people, could affect fetal development.

In 2014, after a couple of observational studies suggested a possible link, the Food and Drug Administration began a formal process to track data on the issue. The findings from the agency’s initial review of the evidence, based on limited and contradictory data, were inconclusive according to a Drug Safety Communication published the following year.

Since then, researchers from several countries, including the U.S., have published a steady stream of observational research. In 2021, an international group of researchers came together to review the evidence and craft a consensus. “We all sat down and said, it’s time to put all this together; we’ve done reviews, there’s more and more evidence,” said lead author Ann Bauer, an epidemiologist at the University of Massachusetts Lowell. “We all felt that it was time for women to have this information.”

Of the 29 observational studies involving 220,000 mother-child pairs, 26 linked prenatal use of acetaminophen to neurodevelopmental disorders including ADHD, autism, language delays, lower IQ, and cerebral palsy among others. Sixteen studies showed a more-pronounced effect with longer-term use of the drug.

By the 1960s, Tylenol was in common use — including by pregnant women — as suggested by this 1967 syndicated column addressing readers’ health and medical questions, published in the Creston News Advertiser. Visual: Illustration by Undark

Bauer pointed out that a handful of observational studies published after the consensus statement also suggest an association.

The group conceded that the observational data is imperfect. The positive association could stem from other factors, such as heredity or the condition that prompted the woman to take the drug. Still, the researchers concluded that the combined weight of the data was strong enough to warrant warning labels on acetaminophen and for health care professional to caution women against indiscriminate use of the drug. Society should act now, they wrote, “not wait unequivocal proof that a chemical is causing harm to our children.”

In science, it’s always possible to find something wrong with individual studies, said David Møbjerg Kristensen, a professor of molecular biology at Roskilde University in Denmark, one of the consensus authors. “But it’s more when you have all the studies lining up that you begin to be concerned,” he said.

But other experts say that it’s misleading to stack up profoundly limited data and conclude that, as a whole, it carries more weight.

The paper from Bauer’s team was “irresponsibly published,” said Nathaniel DeNicola, an obstetrician-gynecologist based in Orange County, California, who helped review the evidence for the American College of Obstetricians and Gynecologists. “It did not reflect the preponderance and overall weight of the data, and it did not reflect the clinical context.”

Support Undark Magazine

Undark is a non-profit, editorially independent magazine covering the complicated and often fractious intersection of science and society. If you would like to help support our journalism, please consider making a donation. All proceeds go directly to Undark’s editorial fund.

DeNicola, who has expertise in environmental exposures and health policy, pointed out that consensus authors didn’t include numerous reviews, including those from major medical organizations, that drew different conclusions. Both ACOG and the Society for Maternal-Fetal Medicine, for example, found no clear evidence that acetaminophen causes fetal developmental issues and no reason to change current medical advice and practice.

In the end, neither did the FDA. In 2018, the agency brought the issue before its Medical Policy and Program Review Council, which provides oversight and direction of policies. The council found the available data didn’t warrant changes on acetaminophen labels or updates to the existing safety communication, wrote FDA press officer Charlie Kohler in an email to Undark.

While the agency continues to monitor the issue, it closed the formal tracking process in 2020 said Kohler, because extensive reviews failed to turn up solid evidence of a link between the drug and neurodevelopmental issues.

The gold standard for understanding the effect of a medication is to randomly assign one group to take the drug and another to get a placebo, with neither researchers nor participants knowing who got what until the end of the study. However, those randomized clinical trials rarely include pregnant women because of potential risks to the fetus. As a result, acetaminophen researchers rely on observational studies and laboratory experiments, including those that test the effects of the drug on experimental animals.

It can be difficult, however, to study neurodevelopmental problems in these animals. For one, researchers wouldn’t diagnose ADHD or autism in a mouse, though some research finds that mice exposed to acetaminophen in the womb are more likely to have problems with learning, memory, motor skills, and social behavior. Additionally, the biological mechanisms that lead to a diagnosis in humans are complex and not well understood, said Kristensen, so researchers don’t know what to exactly look for when they study the brains of laboratory animals.

Research in test tubes and lab animals does show that acetaminophen affects several chemical systems involved in brain development. “The compound is doing multiple different things during development at specific time points when the fetus is vulnerable,” said Kristensen. But whether these factors contribute to neurodevelopmental problems, he added, is unclear. Kristensen said that he expects to publish data within the next year or so that could help clarify the connection.

DeNicola pointed out that consensus authors didn’t include numerous reviews, including those from major medical organizations, that drew different conclusions.

It’s also hard to know what to make of the observational research, which has numerous limitations, according to DeNicola. Many of the studies rely on women’s recall of having taken acetaminophen, which can be faulty weeks and even months later, he said. And instead of a clinical diagnosis of a child’s developmental or behavior disorder, studies often depend on assessments by parents and teachers, which sometimes differ.

One of the biggest issues with the observational research is failing to adequately control for other factors that cause autism and ADHD, particularly heredity, said Per Damkier, head of research in clinical pharmacology at the University of Southern Denmark and a co-author of a consensus rebuttal by the European Network of Teratology Information Services. (Teratology is the study of diseases and conditions that are congenital, or present at birth.) Based on data from Western countries, researchers estimate that up to 80 percent of autism and up to 90 percent of ADHD results from genes children inherited from their parents. If you don’t account for that huge factor, he said, “you inevitably will come up with misleading results.”

For example, a 2017 Pediatrics study included in the original consensus review found that the father’s use of acetaminophen increased a child’s risk of ADHD just as much as the mother’s use during pregnancy. While the researchers theorize that acetaminophen could have altered how the fathers’ genes work, Damkier said that the more likely explanation is that the analysis didn’t sufficiently adjust for heredity.

The pain or illness that prompts a woman to take a pain reliever may also skew the data. In their consensus rebuttal, authors from the Organization of Teratology Information Specialists point to the OTIS’ long-running MotherToBaby study, which found that compared to pregnant women who use acetaminophen for short periods, those who take it longer term report having more mental health conditions, including depression and anxiety, and are more likely to take antidepressants — all factors studies suggest are associated with ADHD and autism in children.

In a 2020 study, epidemiologist Reem Masarwa, then a postdoctoral fellow at McGill University in Montreal, and colleagues explored whether confounding factors such as heredity and maternal illness could be biasing acetaminophen research. The group reanalyzed data from a meta-analysis Masarwa had published two years earlier, which had found a 35 percent increased risk of ADHD in children exposed to prenatal acetaminophen. This time the researchers stringently adjusted for parental ADHD and migraines in the mother.

The ADHD risk nearly disappeared.

The consensus authors cite two studies that overcome the limitation of relying on a mother’s memory of acetaminophen use. The first, a 2019 study, tested umbilical cord blood for traces of acetaminophen and the second, published in 2020, measured acetaminophen in babies’ first bowel movement. Both found that the higher the prenatal exposure to acetaminophen, the greater the chance of a child receiving a physician diagnosis of a neurodevelopmental disorder.

“The beautiful thing about the new studies coming out is that the better the study, the stronger the associations,” said Kristensen. “Suddenly, you can see dose response, which is something we always look for because that argues that there’s something biological going on.”

However, both of those studies have drawbacks as well. Acetaminophen only lingers in the blood for a few hours, so the implication of the cord-blood study is that a single dose right before birth could have a dramatic effect on a child’s brain. Many experts find that result implausible.

Testing a newborn’s bowel movement may be a better measure as it is thought to reflect the mother’s use of acetaminophen during the last two trimesters of pregnancy. But as with some other observational studies, the researchers didn’t account for why the mother took the drug in the first place.

Plaintiff attorney Ashley Keller said that his first responsibility is to help his clients win compensation. In addition, he said, there’s also a public health issue at stake in that the women need full information to make informed decisions.

In April, U.S. District Judge Denise Cote, who is presiding over the pre-trial proceedings, took the unusual step of inviting federal regulators to provide an opinion on whether the science warrants adding a warning to acetaminophen labels about an association between prenatal exposure and ADHD and autism.

The FDA declined to comment on whether the agency would weigh in by the end of July as requested by Judge Cote.

Claiming that FDA regulations and federal laws prevent them from changing Tylenol labels, Johnson & Johnson is pushing for a dismissal. The company continues to evaluate the science and has not seen any evidence that acetaminophen use during pregnancy causes fetal development issues, Witt, the company spokesperson, wrote in an email. “Leading medical organizations agree with the current product label which clearly states, ‘If pregnant or breast-feeding, ask a health professional before use.’”

“As far as the epidemiology, there’s a pretty perfect study that could be done,” Bauer said, pointing out that, with smartphones, women can easily record what they take and why.

Bauer said the lawsuits are helping get the word out about the accumulating research on acetaminophen’s risks. Up until now, she said, a lot of women have viewed the drug as “completely innocuous,” something to take “whenever they are in discomfort.” Others disagree. DeNicola said that the women he sees in his practice are conscientious about using acetaminophen.

One thing everyone agrees on is the recommendations for acetaminophen use, said DeNicola: “Use it only where indicated in the lowest dose for the shortest duration.” So, for pain or fever that means taking one pill, one time, to see if symptoms ease, he said.

Another point of real consensus is the need for more research. Bauer said that several groups are looking at possible mechanisms for how acetaminophen could affect development. “As far as the epidemiology, there’s a pretty perfect study that could be done,” she said, pointing out that, with smartphones, women can easily record what they take and why. “But it’s going to take us many, many years is the problem.”

“Something that is potentially this important for, not just fetal, but for childhood brain development should be studied,” DeNicola said. “And it should be studied in a real way.” For example, he would like to see studies that use blood tests and other measures throughout pregnancy to accurately track exposures not just to acetaminophen, but also to environmental substances of greater concern to fetal development, such as industrial chemicals. And, since a baby’s brain continues to develop after birth, studies should account for use of acetaminophen in childhood — a glaring omission from most of the current research, said DeNicola.

At the moment, no one is conducting exactly that type of study on acetaminophen.

Vintage Tylenol packaging and advertisements — including some dating back to the original product for children marketed by McNeil Laboratories in the 1950s — are now sold as collectors items online. Visual: Illustration by Undark

However, FDA spokesperson Charlie Kohler said that agency is conducting a small study to see how people’s bodies absorb, metabolize, and excrete the drug. The agency is also planning a toxicology study in 2024, pending approval of funding.

And Chambers said that she is coordinating a study funded by the National Institutes of Health of about 8,000 mother-child pairs in 25 sites across the U.S. The Healthy Brain and Child Development Study will capture information on prenatal exposures, including to acetaminophen, and use brain imaging and standardized assessments to track brain development in children. Researchers will release data annually, so information on prenatal exposure to acetaminophen will begin to emerge in the next couple of years, she said.

If there’s anything good to come out of the controversy, it’s that every drug, over the counter or not, deserves rigorous research on safety, said Chambers — and society hasn’t invested in systematically doing that type of research. She pointed out that while the FDA requires new drugs to be assessed for safety during pregnancy, nobody has the resources of motivation to do that for old drugs like acetaminophen.

“And we need to stop that,” said Chambers. “We need to turn that wheel around, pay attention and do the work that needs to be done.”

UPDATE: A previous version of this piece incorrectly described the academic affiliation of Christina Chambers. She is at the University of California, San Diego, not the University of San Diego.

Teresa Carr is a Colorado-based investigative journalist and the author of Undark's Matters of Fact column.