During an April 17 congressional hearing, Health and Human Services Secretary Robert F. Kennedy Jr. called for retraction of a new Danish study that didn’t find a link between Tylenol and autism, repeatedly calling it “garbage” and baselessly suggesting that it was industry-generated and “fraudulent.”
There is no evidence of fraud or industry involvement, and the criticism Kennedy made was a limitation the authors of the paper acknowledged — not legitimate grounds for retraction, according to scientists.
Beginning with a press conference about autism in September — the Kennedy-imposed deadline for knowing the cause of the “autism epidemic” — President Donald Trump has repeatedly told pregnant women not to take Tylenol unless “absolutely necessary.” Kennedy has been a bit more circumspect on the topic, speaking of a “potential association” between prenatal Tylenol, also known as acetaminophen, and later autism diagnoses in children and calling the literature finding a connection “very suggestive.”
As we wrote in September, some studies have shown an association between prenatal acetaminophen use and autism. However, experts told us that these associations were likely not causal, and instead probably due to traits shared among people who tend to take more acetaminophen in pregnancy, such as a hereditary susceptibility to autism.
The new Danish study, published April 13 in JAMA Pediatrics, looked at national prescription fulfillment records for mothers of more than 1.5 million children and corresponding health records, finding no association between taking acetaminophen or taking greater doses of the drug during pregnancy and later autism diagnoses in the children.
When asked about the Danish study at the House Education and Workforce Committee hearing on April 17, Kennedy moved to discredit it. “The study is a garbage study. It should be retracted,” he told Republican Rep. Virginia Foxx of North Carolina. Kennedy went on to criticize the study for relying on prescription records when acetaminophen is also available over the counter. “It was a garbage in, garbage out study,” Kennedy continued. “The industry has the capacity to generate these studies all the time, and it’s fraudulent. It should be retracted.”
The study did rely on prescription data, which can lead to incomplete data on the use of the drug, Dr. Kira Philipsen Prahm, a doctor in the Center for Fetal Medicine at the Copenhagen University Hospital Rigshospitalet and first author of the study, told us via email. But such a limitation “does not automatically invalidate results,” she said. “The key question is whether the misclassification is likely to meaningfully bias the findings.” Her team’s analyses, along with prior research, indicate that “if there were a strong causal effect” of acetaminophen on autism, “it would be unlikely to be entirely obscured by this limitation,” she said.
Brian Lee, a professor of epidemiology at Drexel University’s Dornsife School of Public Health, told us that most acetaminophen is prescribed in Denmark, following restrictions on how much of the medication can be sold without a prescription. This makes Denmark a relatively good location to do a prescription-based study, he said, contrary to Kennedy’s implication that the approach invalidated the study. These restrictions were in place during the latter years of the study.
Furthermore, Prahm said, her team’s study did not find “a pattern suggesting increased risk with greater recorded exposure.” If acetaminophen were causing autism, one would expect to see more cases with increasing doses.
Nor are papers retracted simply because they have limitations, which all studies have. Prahm and her colleagues wrote in their paper that information about individuals’ over-the-counter acetaminophen use was missing and that “thus, the true exposure level among those with low-level exposure was likely underestimated,” while also explaining why they thought this was unlikely to have introduced meaningful bias.
Kennedy has a history of trying to “wield his considerable influence” to “force a retraction of a study without a legitimate reason,” Lee said, referring to a study about a common vaccine ingredient Kennedy said last summer should be retracted.
Legitimate reasons for retraction, Lee said, would include “analytical errors that affect the qualitative conclusions of the study, integrity issues, or loss in confidence of findings by the authors.” Prahm’s study “does not appear to feature any of these issues,” he said, calling Kennedy’s calls for retraction “unwarranted and politically coercive.” Lee was co-author of a 2024 Swedish study that pointed away from a causal association between prenatal acetaminophen use and autism in children, but he was not involved in the new Danish study.
Dr. Per Damkier, a professor in the department of clinical research at the University of Southern Denmark, told us via email that Kennedy “is well outside his domain of expertise” in assessing the scientific merits of the study. Damkier was not involved in the new study but has studied acetaminophen use during pregnancy.
Prahm said that the study was “conducted using nationwide Danish registry data and the pharmaceutical industry was not involved in funding or any other part of the study.” The study lists Danish governmental and hospital funding. One of the nine authors disclosed funding by a pharmaceutical company for unrelated work evaluating a contraceptive pill.
HHS did not reply to a request asking for the basis for Kennedy’s claims about the Danish study.
Missing Context on Acetaminophen in Denmark
Kennedy faulted the Danish study for using prescription data and for the low percentage of women it recorded as using acetaminophen. “Only 2% of the people in this study got Tylenol during pregnancy, according to the endpoint,” Kennedy told lawmakers. “In fact, we know, because Tylenol is available by over the counter, most of you have taken Tylenol. Very few of you have ever gotten a prescription.”
But Kennedy was missing context on acetaminophen in Denmark, which has been increasingly obtained via prescription in recent years.
“Reliance on prescription records alone would be bad in a setting like the US, where most acetaminophen use is” over the counter, Lee said. “However, Denmark is not the US.”
Damkier said that before 2014, “more than 60% of all acetaminophen sold in Denmark” was over the counter. But in late 2013, Denmark limited the quantity of acetaminophen that could be sold without a prescription. Following this change, “more than 80% of acetaminophen sold has been prescription based,” he said, citing his own research on the topic. “I believe exposure data from 2014 and onwards are valid and representative with low risk” of misclassifying acetaminophen use, he said.
The new study looked at prescription records from pregnancies for children born between 1997 to 2022. Damkier said that the study “can be criticized” for using prescription data prior to the change in prescription regulations but that he believes “the conclusions of the authors are substantiated” overall. “By and large, this large population-wide study supports the findings from the most recent studies: Exposure to acetaminophen during pregnancy is not associated with an increased risk of childhood” autism, he said.
Prahm said that she and her co-authors had done further analyses to see if the findings varied before or after 2013, but the team “found no statistical differences between the two periods.”
Kennedy also provided a relatively high-end estimate for acetaminophen use during pregnancy in Denmark. “Fifty percent of the women in Denmark, we know from other studies, actually took Tylenol during pregnancy,” Kennedy said. “So the study was comparing people, women who took Tylenol during pregnancy to women who took Tylenol during pregnancy.”
HHS did not reply to a question about where Kennedy got this statistic, but older, self-reported data from the Danish National Birth Cohort found this relatively high rate of use. Estimates of acetaminophen use during pregnancy vary, and one more recent study found that 6% of women reported using the medication during the first trimester.
Lee said that many women in the Danish National Birth Cohort study were missing responses on acetaminophen use and were not included, saying that “the 50% is almost assuredly an overestimate.”
Furthermore, Lee and Prahm both objected to Kennedy’s characterization of the new study as comparing “women who took Tylenol during pregnancy to women who took Tylenol during pregnancy.”
“That is not an accurate description of the study design,” Prahm said. “While some individuals classified as unexposed may in fact have used over-the-counter acetaminophen, this does not mean the two groups are equivalent.”
Pros and Cons of Prescription Data
The Danish study is not alone in using prescription data. Lee explained that using prescription data has “advantages and disadvantages.” An advantage is that it provides an objective record of drug supply, whereas studying over-the-counter exposure requires asking people to report on their own use, he said.
People can misreport their medication use, Prahm said, or the data can be influenced by recall bias, a phenomenon where people can remember things differently depending on later events. For example, a parent with a child diagnosed with autism might remember their medication use during pregnancy differently than a parent without this experience.
Furthermore, while prescription-based studies do miss some exposures to acetaminophen, they are likely to capture the most impactful use.
“Prescription based exposure likely captures those women who use substantial amounts of acetaminophen as opposed to [over-the-counter] based use, which tends to be low and sporadic,” Damkier said. “If there is no signal for prescription-based use, it is consequently exceedingly unlikely that sporadic [over-the-counter] use be associated with an increased risk” of autism.
Regardless, researchers don’t rely on single studies to draw conclusions. Rather, they look for a pattern of replication among studies done using various methods and datasets, David S. Mandell, a psychiatry professor at the University of Pennsylvania Perelman School of Medicine and director of the Center for Mental Health Policy and Services Research, told us via email. “When we see replication, we grow more confident in the findings.”
Multiple studies have found that associations between prenatal acetaminophen use and neurodevelopmental conditions go away when comparing siblings. In recent months, two review studies have pulled together the available data, concluding that the evidence does not show any clear or “clinically important” link between prenatal exposure to the medication and autism.
“We now have studies from Nordic countries, Japan and Taiwan showing that Tylenol doesn’t cause autism,” Mandell said. The degree of acetaminophen use varied in the studies, “and it doesn’t make a difference in the findings.”
Prahm emphasized that her team aimed to “contribute one piece of evidence” to be interpreted in the context of the broader literature. “Overall, the current evidence does not establish a clear association,” she said.
Editor’s note: FactCheck.org does not accept advertising. We rely on grants and individual donations from people like you. Please consider a donation. Credit card donations may be made through our “Donate” page. If you prefer to give by check, send to: FactCheck.org, Annenberg Public Policy Center, P.O. Box 58100, Philadelphia, PA 19102.
Source link
