Could Tylenol, one of the most ubiquitous medicines in the U.S., be linked to some cases of autism spectrum disorder when given during pregnancy? And could leucovorin, a form of folic acid, improve the verbal ability of some people with autism spectrum disorder?
Those are two possibilities that President Trump and Robert F. Kennedy Jr., the secretary of health and human services, seized on Monday in an announcement at the White House. Neither is new; data suggesting that Tylenol use in pregnancy could increase the risk of an infant developing autism and that leucovorin may help some patients have circulated for more than a decade. But in both cases it is exceedingly difficult to tease out verifiable risks and benefits from hypotheses alone.
It’s also important to remember that the exact definition of autism spectrum disorder has varied over time, and that the term is used to describe a wide range of conditions, affecting everyone from those who need round-the-clock care to successful white-collar professionals. Not everyone necessarily reacts the same way to the same medications.
Most researchers who have looked at the question of whether Tylenol taken during pregnancy is tied to autism do not claim to know that there is a relationship between the two, and some prominent researchers now think the common painkiller is unlikely to play a role.
But there is substantial disagreement. A review of 46 studies published in the journal Environmental Health just last month and co-authored by researchers at Harvard University and the Icahn School of Medicine at Mount Sinai concluded that although the data are mixed, it is worth warning pregnant people about the potential risk of taking Tylenol.
Advertisement
“Appropriate and immediate steps should be taken to advise pregnant women to limit acetaminophen consumption to protect their offspring’s neurodevelopment,” the paper concluded.
One of the authors of that paper, Andrea Baccarelli, dean of the faculty at Harvard T.H. Chan School of Public Health, was an expert witness for a plaintiff in a case involving potential links between the use of acetaminophen, the active ingredient in Tylenol, during pregnancy and neurodevelopmental disorders.
Causation, or just correlation?
Multiple studies using different methodologies show that mothers who take Tylenol during pregnancy are more likely to give birth to children who are later diagnosed with autism spectrum disorder. Researchers have been clever in studying the relationship, looking at everything from mothers’ memory of how often they took acetaminophen to levels of chemicals related to Tylenol in umbilical cord blood.
Advertisement
The relationship is relatively small — one study says 5% increase in risk, though a meta-analysis of other studies put the increase in risk at 20% — and would not explain a substantial increase in autism cases in recent years. (Many, though not all, researchers argue that increase is the result of widened criteria for autism spectrum disorder.) But what researchers debate is whether Tylenol might cause autism, or whether Tylenol is simply more often used by people who experience certain conditions during pregnancy, such as infections or migraines, which might also be linked to autism.
This is a key problem in science. Ice cream consumption increases in the summer, as do sunburns and shark attacks. But ice cream does not cause sunburns or shark attacks — they all just happen more often during the summer.
When it comes to Tylenol and autism, one notable paper, published in JAMA last year, tried to deal with this issue by looking at cases in which a mother took acetaminophen in one pregnancy, but not another. The idea is that if genetic or other environmental factors were leading to an increase in autism risk, both siblings would be exposed to the same risk.
The study used a database of 2.5 million children born in Sweden from 1995 through 2019. Of those children, about 186,000 had been exposed to Tylenol in utero.
“In our opinion, this suggests that associations that we see between acetaminophen use during pregnancy and autism or ADHD are most likely non causal and probably due to familial confounding like genetics,” said Brian Lee, a co-author of the study and a professor of epidemiology at Drexel University.
If a mother has a higher genetic risk for ADHD or autism, Lee said, that person is also likely to have more pregnancy pain, more migraines, and use more pain medications — a finding he says has been replicated across the world from the U.K. to Norway to Japan.
“This isn’t something that’s like, oh, you know, we’re only finding this in one sample,” he said. Instead, a genetic risk for autism and ADHD is leading to an association between acetaminophen and those conditions.
Advertisement
This makes sense because many scientists believe that autism spectrum disorder is the result of genetics, and in most people not of a single genetic mutation but of a combination of genetic changes that, depending on how they interact, can cause any number of cognitive or behavioral issues.
Lee noted that it is certainly possible for a drug that is given during pregnancy to increase the risk of autism spectrum disorder, and that the relationship between one epilepsy medicine, valproate, and autism is so strong that researchers use the medicine to cause autism-like behaviors in rats.
Criticism of the Swedish sibling paper
But not everyone agrees that the sibling paper from JAMA is ironclad, and the review of 46 studies includes this paper.
Ann Bauer, a researcher at the University of Massachusetts, Lowell, and a co-author of the review published in Environmental Health, told STAT in an interview that one major issue with the Swedish analysis is that use of acetaminophen was incredibly low. Usually, half of mothers report using the drug at some point during pregnancy; in this database it was only 7.5%. That leads to a concern that mothers were taking acetaminophen but did not report it.
That, she and her co-authors argue, leads to a problem with sibling studies, which they say can underestimate risks in cases where data are not captured. The authors still favor warning pregnant people about the risk, especially if they do not have a high fever and might be able to avoid using it.
The problem, however, is that recommending pregnant people not use acetaminophen leaves them few other options at a time when other drugs can pose clearer risks to a fetus.
The United Kingdom’s Medication and Healthcare products Regulatory Agency said in a statement Monday that “based on rigorous assessment of the best available scientific evidence” it still recommends acetaminophen, known as paracetamol in the U.K., during pregnancy when used as directed.
“Untreated pain and fever can pose risks to the unborn baby, so it is important to manage these symptoms with the recommended treatment,” the agency said.
Advertisement
Does leucovorin help people with autism?
The case of leucovorin is simpler to explain: There are several studies that indicate the medicine can improve the performance of people with autism on standardized tests used to measure verbal ability. But these studies are quite small.
One study, published in the European Journal of Pediatrics, followed 80 patients between the ages of 2 and 10 who were randomly assigned to receive either leucovorin or a placebo. Neither their families nor their doctors knew who received the treatment. At the end of 24 weeks, the children who received leucovorin scored 1.2 points higher on a 60-point scale used to measure autism severity than the children who did not.
The result was statistically significant, but smaller studies are prone to false positive results. The normal course for researchers before making a recommendation would be to conduct a much larger randomized controlled trial to verify that the treatment is beneficial. Most such trials include hundreds or thousands of patients, and provide the best evidence of efficacy and safety.
For both acetaminophen and leucovorin, another problem exists: Generally, high-quality studies of medicines are done by pushing manufacturers to test them. But even though Tylenol, the leading brand of acetaminophen, is a big seller, both acetaminophen and leucovorin are available as generics, which means companies are less likely to pay for large observational studies or randomized controlled clinical trials.