Though there is some research supporting President Trump's claim linking Tylenol use to autism spectrum disorder, Dr. Joel Shulkin argues that these studies are insufficient to demonstrate any causative link.
On Sept. 22, 2025, President Donald Trump, alongside Health and Human Services Secretary Robert F. Kennedy, Jr., announced that the US Food and Drug Administration (FDA) would be notifying physicians that the use of acetaminophen (Tylenol) by expectant mothers can be associated with a "very increased risk of autism." This announcement has been met with widespread criticism from the scientific community.
Autism is a neurodevelopmental disorder (NDD) characterized by difficulties with social interactions, verbal and non-verbal interactions, and repetitive, characteristic patterns of behavior. It is also known as Autism Spectrum Disorder (ASD) because those diagnosed with ASD exhibit a wide range of symptoms. For decades, the exact cause of autism has not been known due to the inherent complexity of the disorder, with a range of potential environmental and genetic causes. This announcement made on the causal link stirred widespread concern. Perhaps this is a déjà vu of the Wakefield report (false report linking autism to MMR vaccine). The claim just didn't sound... well, sound.
In an email to The News-Letter, Dr. Joel Shulkin, an assistant professor of pediatrics and pediatrician in Pediatric Developmental Medicine at Hopkins All Children's Hospital, criticized the reported correlation between Tylenol and autism.
"Observing that a hurricane occurred on the same day as the Super Bowl does not mean the Super Bowl caused the hurricane (or vice versa)," he wrote.
Prior research studies have indeed shown a correlation between acetaminophen and autism. A notable one is a study that found an association between MMR vaccination followed by Tylenol use and autism. However, while such studies demonstrate a correlation, Shulkin stated that they cannot determine causation.
Additionally, a February 2025 meta-analysis showed that most studies reporting correlations between Tylenol use and autism exhibited high selection bias, variability in selection, adjustment for potential confounders and unmeasured familial confounding. In addition, a major study published last year conducted on about 2.5 million Swedish children, including sibling case-controls, found no association of acetaminophen use during pregnancy with children's risk of autism.
"Therefore, there is no evidence at this time that prenatal acetaminophen causes autism," Shulkin said.
It is also worth noting that claims linking acetaminophen use during pregnancy to brain injury, autism and other neurodevelopmental disorders in children are mainly based on theoretical and animal models. Shulkin mentioned a study whose "evidence" of causation was based on 20 animal models and theoretical observations. A 2025 systematic review suggested an association between acetaminophen exposure during pregnancy and increased incidence of NDDs. Again, association, not causation. It is important to note that one of the authors, Andrea Baccarelli, of this study served as a paid expert witness in a lawsuit concerning prenatal acetaminophen use and neurodevelopment disorders in offspring, an acknowledged conflict of interest.
Lastly, a study published in 2019 suggested that higher acetaminophen levels in cord plasma were associated with higher childhood risks of ASD or attention-deficit/hyperactivity disorder (ADHD). According to Shulkin, there are two things worth noting about this study. One, this study was conducted from 1998 to 2018. The Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5) was released in 2013, introducing a significant change in the diagnostic criteria for autism. Before 2013, due to the use of older classification systems, some cases may have been omitted, potentially introducing some uncertainty. Second, the authors did acknowledge that they were unable to exclude potential residual confounders due to the observational nature of the study, and mentioned unmeasured genetic and environmental factors as their reasons.
President Trump's announcement made it seem that Tylenol is the sole cause of autism. Shulkin said that several members of the administration indeed tried to make that claim.
"This, of course, is nonsensical, as many women do not use any pain relievers during pregnancy and yet still have children who are autistic, while other women use acetaminophen and their children are not autistic," he wrote. "Given the widespread use of acetaminophen in pregnancy and early childhood, if use at appropriate dosages and intervals is causative of autism, the prevalence should, in fact, be much higher than it is."
For a long time now, acetaminophen has been known to be the safest pain and fever medication for expectant mothers. According to Shulkin, it is apparent from the announcement that pregnant mothers were being discouraged from taking this medication, and no alternative solution to that medication was given. He also mentioned that untreated fever and pain during pregnancy are associated with a high risk of perinatal complications and can contribute to a higher likelihood of postpartum depression.
"The American Academy of Pediatrics, American Psychiatric Association, American College of Obstetricians & Gynecologists and many other medical organizations have all issued statements within the past month emphasizing the safety of acetaminophen use during pregnancy and childhood," Shulkin explained. "Therefore, my advice to expectant parents is to listen to the advice of your physician, not to that of a bureaucrat."
The announcement trivialized autism and decades of rigorous research into such a complex NDD. This announcement rushes a process that science is taking with caution and precision. But Shulkin concluded with a strong parting shot that offers perspective as scientists, physicians and families continue to wrap their heads around this somewhat enigmatic condition.
"From my perspective, the research should be directed not on 'finding the cause' or 'finding a cure' but identifying how to improve early provision of services, to train and support parents to promote their child's development and to eliminate disparities in care across the lifespan," Shulkin argued.