Acetaminophen (Tylenol) has been considered the safest option for pain and fever in pregnancy.
But a new systematic review from Mount Sinai and Harvard raised concerns: could prenatal use increase the risk of autism or ADHD?
The conversation escalated when President Trump urged pregnant women to “fight like hell” against taking acetaminophen. But what does the evidence actually show?
Episode Highlights
- The Headlines — how a new study and political commentary ignited debate
- Acetaminophen’s Role — why it became the first-line fever and pain option in pregnancy
- Autism & ADHD — what these conditions are, and why reported diagnoses are rising
- The Evidence — 46 studies, including prospective cohorts, cord blood data, and sibling comparisons
- Association vs. Causation — why links don’t prove acetaminophen is the cause
- The Contradiction — large sibling studies showing no increased risk
- Biological Plausibility — what animal models suggest, and their limitations
- Regulators’ View — why WHO and others still recommend acetaminophen
- Practical Takeaways — how to balance risks, relief, and real-world pregnancy needs
Show Notes
The Study at a Glance
A systematic review published in Environmental Health (Aug 2025) analyzed 46 observational studies on prenatal acetaminophen exposure. Results varied widely — some showed increased risk of autism or ADHD, some showed no link, and a few even suggested protective effects.
Sibling comparison studies from Sweden and Norway, involving millions of births, found no increased risk, contradicting much of the other data.
Why This Matters
Observational studies can show associations, but they cannot prove causation. Relative risk increases (like 30% higher risk of ADHD) often sound dramatic, but the absolute risk is small, often less than a 1% increase.
Meanwhile, untreated fever and severe pain in pregnancy carry their own risks for mother and baby.
Practical Takeaway
Most experts, including the WHO, continue to recommend acetaminophen as the safest option during pregnancy when used at the lowest effective dose, for the shortest time possible.
If you’re pregnant or supporting someone who is, the key is making informed decisions, not fear-driven ones.
For more health insights, subscribe to The Dr. Kumar Discovery Podcast on any major platform.
Cheers!
Transcript
[00:00 –> 00:19] On this episode of the doctor Kumar discovery. Autism in particular has been a lightning rod issue. It is very divisive. No parent wants to hear that their child has autism because it can set them apart and create real challenges in life. So when people point fingers at possible causes, fear and sensationalism tend to follow.
[00:19 –> 00:42] And that’s important here because the paper didn’t and could not look at randomized clinical trials of acetaminophen use in pregnancy because none exist. If you’re pregnant and suffering with severe pain or headaches, or if you have a fever, should you just push through and deny yourself treatment because of this new analysis? My answer would be no. My name is Doctor. Ravi Kumar.
[00:43 –> 01:11] I’m a neurosurgeon in search of the causes of human illness and the solutions that help us heal and thrive. I want you to join me on a journey of discovery as I turn over every stone in search of the roots of disease and the mysteries of our resilience. The human body is a mysterious and miraculous machine with an amazing ability to self heal. Let us question everything and discover our true potentials. Welcome to the Doctor.
[01:11 –> 01:19] Kumar discovery. Welcome to the Doctor. Kumar discovery podcast. My name is Doctor. Ravi Kumar.
[01:19 –> 01:55] I’m a board certified neurosurgeon and assistant professor at UNC. Today, we’re gonna discuss a study published this past August by researchers at Mount Sinai and Harvard. It suggested an associative link between acetaminophen use during pregnancy and a diagnosis of autism or ADHD in children. So why bring this up now? Because on September 22, president Trump gave an address to the nation declaring that acetaminophen use in pregnancy is associated with autism and ADHD, and he urged women to, quote, unquote, fight like hell not to take it.
[01:55 –> 02:22] That statement is a lightning rod. When a major political figure makes sweeping scientific claims and medical recommendations, alarm bells should go off. It’s not that politicians can’t care about science, but medicine is not their lane, and they often have other motivations. And in this case, the president was referencing data on a hot button topic that most of us have never seen and that he likely does not understand. So how should we approach this?
[02:22 –> 03:12] Well, the most reasonable thing to do here is to look at the study directly, read the evidence, and understand what it actually shows, and then make up our minds based on that data. What we shouldn’t do is jump up and proclaim that what the president said was right or wrong because that immediately puts us on a team where our affiliation to the team matters more than the actual truth. So here’s what I wanna have by the end of this episode, a clear, unbiased understanding of the study that president Trump referenced. We’ll walk through what it reported, what the implications might be, where the strengths and weaknesses lie, and how to apply it to real life decisions. And if you’re a pregnant woman or their loved one, I want you to leave this episode with the information you need to make decisions without fear and armed with knowledge.
[03:12 –> 03:28] That means understanding not only the potential risks, but also the benefits and the probabilities behind them. So that’s what you’ll walk away with by the end of this podcast. Okay. Before we dive in, I just wanna give you a quick disclaimer. I’m a doctor, but I’m not your doctor.
[03:28 –> 03:53] This podcast is for informational purposes only. It’s not intended to diagnose, treat, or replace personalized medical advice. My goal here is to give you clear information that helps you think critically about health and medicine and to break down complex topics so you feel better equipped to make informed decisions. And one more note, this podcast is entirely separate from my role as assistant professor at UNC. Okay.
[03:53 –> 04:15] So before we start analyzing the study, let’s step back and ask what exactly is acetaminophen. Most people know it by the brand name Tylenol. Around the world, it’s called paracetamol. Chemically, acetaminophen is a small compound that is an extraordinarily effective analgesic and antipyretic. In other words, pain killer and fever medication.
[04:15 –> 04:50] It was first derived from coal tar, which is a byproduct of processing coal. Back in the day, researchers noticed that a substance in coal tar called acetanilide could lower fever, but there was one major problem. It was highly toxic. What they eventually discovered was that acetanilide was metabolized in the body into a safer compound called acetaminophen, and it was acetaminophen itself that was having biological effects, like reducing fever and relieving pain, and it did it without the same level of toxicity. From there, acetaminophen was developed and later marketed as Tylenol.
[04:50 –> 05:16] In medicine, it became one of our most trusted tools. It was a safe and reliable analgesic and antipyretic, meaning it lowers fever and pain. Pregnancy is where acetaminophen has been extraordinarily helpful. Other common pain and fever reducers like NSAIDs, such as ibuprofen or naproxen, can cause problems for the fetus and are generally considered unsafe during pregnancy. That left acetaminophen as a linchpin medication.
[05:16 –> 05:41] It was one of the few safe and effective options doctors could use to help women manage pain and fever while pregnant. Its use in pregnancy is now widespread. More than half of pregnant women around the world take acetaminophen at some point during their pregnancies, making it the first line choice for pain and fever in this population of women. Okay. So the next thing we need to answer is what exactly is ADHD and autism?
[05:41 –> 06:11] These are the two neurodevelopmental disorders at the center of the study that president Trump referenced. ADHD or attention deficit hyperactivity disorder is defined by persistent inattention, hyperactivity, and impulsivity. The symptoms have to be significant enough to interfere with daily life. And according to the DSM five criteria, which is the standard psychiatric manual, they must begin before the age of 12. Autism or autism spectrum disorder is also a neurodevelopmental condition.
[06:11 –> 06:39] It’s characterized by ongoing difficulties in social communication and interaction along with restrictive or repetitive behaviors and interests. These symptoms can start early in life and cause real functional challenges. The severity can vary widely, which is why it’s called a spectrum. When people say someone is on the spectrum, they’re referring to this broad range of possible presentations. Now one of the concerns in society is that autism seems to be on the rise.
[06:39 –> 07:16] According to the CDC, in 02/2018, the prevalence was one in forty four children. By 2020, it was one in thirty six, and by 2022, one in thirty one, which is an incidence of about three point two percent. At first glance, that looks like a dramatic increase, but the CDC also notes that much of this rise may be due to better identification, earlier evaluations, improved documentation, and greater access to services. In other words, the numbers may reflect better recognition rather than a true biological surge. Still, we can’t say with certainty whether improved detection is the only factor at play.
[07:16 –> 07:42] ADHD shows a different pattern. Between 2016 and 02/2022, diagnoses rose from about ten point five percent to eleven point four percent, which is a smaller increase, but experts also attribute this to improved awareness and identification. Though again, we can’t completely rule out biological or environmental conditions. Autism in particular has been a lightning rod issue. It is very divisive.
[07:42 –> 08:10] No parent wants to hear that their child has autism because it can set them apart and create real challenges in life. So when people point fingers at possible causes, fear and sensationalism tend to follow. The most notorious example is the 1998 case series by British surgeon Andrew Wakefield. He published data on just 12 children, claiming the MMR vaccine, that’s measles, mumps, rubella, might trigger autism. The media seized on it, and panic spread.
[08:10 –> 08:42] But later, investigative journalist Brian Deer uncovered that Wakefield had been paid by lawyers suing vaccine makers and had manipulated his data. Wakefield eventually lost his medical license, and The Lancet, had published his paper, fully retracted it. The BMJ later called it an elaborate fraud. Meanwhile, dozens of large, well conducted studies have shown no link between vaccine and autism. Yet here we are, more than twenty five years later, and some people still fear vaccines because of that one fraudulent report.
[08:43 –> 09:18] The truth is no study has ever proven a causal connection between vaccines and autism. Vaccines remain one of the most powerful tools we have for reducing childhood illness and death, and their benefits far outweigh the risks. But Wakefield’s study left a lasting scar in the public’s mind. It shows how one sensational claim, even if retracted, can shape fear and mistrust for decades. And that’s an important backdrop as we think about new claims being made today, like the current announcement of a link between acetaminophen use in pregnancy and neurodevelopmental disorders.
[09:18 –> 10:14] So that brings us to the study at the center of this discussion, which is drawing an association between prenatal acetaminophen use and autism and ADHD. It was published on 08/14/2025 in Environmental Health, which is a BMC open access journal. The paper titled Evaluation of the Evidence on Acetaminophen Use and Neurodevelopmental Disorders Using the Navigation Guide Methodology is a systematic review of studies examining links between prenatal acetaminophen exposure and ADHD in autism. The authors opted for a qualitative synthesis rather than a meta analysis, which means they graded study quality and summarized the direction and strength of the evidence instead of primarily pooling effect sizes. And that’s important here because the paper didn’t and could not look at randomized clinical trials of acetaminophen use in pregnancy because none exist.
[10:14 –> 10:51] What we have instead are observational studies, and by their nature, those are weaker. Observational data can show associations but cannot prove causation. That’s where the navigation guide comes in. By scoring and ranking these observational studies, the authors try to give more weight to stronger, better designed ones, and less to the weaker, more biased ones, in an attempt to extract the clearest possible signal from evidence that is by default weak and heterogeneous. This study pulled together 46 observational studies on prenatal acetaminophen use and its association with ADHD and autism.
[10:51 –> 11:19] The results were all over the map. Out of the 46 studies, 27 showed a positive association, meaning that mothers who reported acetaminophen use during pregnancy were more likely to have children later diagnosed with neurodevelopmental disorders. Nine of the studies showed no link at all. And interestingly, four even suggested a protective effect, where acetaminophen use was actually linked to a lower risk of ADHD or autism. Most of the studies included were perspective and design.
[11:19 –> 11:42] In other words, researchers followed pregnant women over time, asked them to self report their acetaminophen use, and then track their children as they grew to see whether they developed ADHD or autism. So here’s where the limitations come in. First, self reporting is never perfect. Mothers may forget how much they took, when they took it, or even whether they took it at all. That introduces error right from the start.
[11:42 –> 12:00] Second, and this is the bigger issue, even when you see an association, it doesn’t prove causation. Here’s an example. There’s an association between ice cream sales and drowning deaths. Studies show that they go up together, but ice cream doesn’t cause drowning. The real factor is heat.
[12:00 –> 12:36] When it’s hot, more people buy ice cream and more people swim, so unfortunately, drowning deaths increase. That’s a perfect example of association without causation. So in this research, just because a mother took acetaminophen during pregnancy and her child later developed ADHD or autism doesn’t mean the medication caused it. It could just as easily reflect other factors. Maybe there’s a genetic predisposition or lifestyle or environmental exposures which made her both more likely to need acetaminophen and also more likely independently to have a child with neurodevelopmental disorder.
[12:36 –> 13:08] That’s the challenge with these kinds of observational studies. They can show patterns, but they can’t untangle whether acetaminophen itself is responsible or whether it’s simply a marker for other risks that are already present. Now beyond prospective studies, the researchers also looked at biomarker studies. These measured acetaminophen and its metabolites directly in cord blood and in meconium. Cord blood is the blood taken from the umbilical cord right after birth, and meconium is essentially the baby’s first stool, sometimes passed within the placenta.
[13:08 –> 13:40] By measuring acetaminophen levels in these samples, the researchers weren’t relying on a mother’s memory of what she took. They could actually see biochemical evidence of exposure in the baby. Then by following those children over time, they could track who did or did not go on to develop neurodevelopmental disorders like ADHD or autism. They also analyzed sibling comparison cohorts. In this type of study, children from the same family are compared to one another, so the genetics, the household environment, and many of the social and lifestyle factors are held constant.
[13:40 –> 14:24] If a mother used acetaminophen during one pregnancy but not during another, the children could be compared directly, which helps control for many confounding variables unrelated to acetaminophen itself. Finally, the review included animal studies and in vitro mechanistic work. These don’t prove cause and effect in humans, but they can show whether there’s a biologically plausible pathway by which acetaminophen might influence neurodevelopment. So these were the types of evidence considered, perspective and case control studies, biomarker studies, sibling comparisons, and mechanistic models. Each comes with its own strengths and limitations, but together, they form a body of research this review tried to analyze and come up with a conclusion.
[14:24 –> 14:51] Now, do you remember me saying that this was a qualitative study? Many of the studies they looked at here had odds ratios associated with them, meaning they had relative risk of the association between acetaminophen and ADHD or autism. But this study wasn’t focusing on quantitative data. It was more focused on qualitative, meaning they wanted to know what was the overall direction and quality of these studies. They were focusing on this question.
[14:51 –> 15:29] Is the association between prenatal acetaminophen use in autism and ADHD valid? But because we all wanna know numbers, I’m gonna just put out some numbers of the studies they reviewed here so you can get an idea of the kind of data that this paper was dealing with. For ADHD, probably the largest study showed a hazards ratio of one point two nine to one point three seven for prenatal use of acetaminophen in ADHD. That means if a woman used acetaminophen in pregnancy, there was a twenty nine to thirty seven percent increase in relative risk of ADHD. Now remember that relative risk is very different than absolute risk.
[15:29 –> 15:58] So if the baseline risk of ADHD is eleven point four percent in the population, a thirty percent relative risk increase would be about three point four percent absolute risk increase, meaning three point four percent more kids would get ADHD based on these stats. Okay. Now let’s look at autism. A pooled meta analysis study this paper looked at showed an increase of one point one nine times or about nineteen percent. The baseline prevalence of autism in children is three point two two percent.
[15:58 –> 16:25] If you increase that by nineteen percent, the rate rises to about three point eight three percent. So the absolute risk increase is only about zero point six percent for acetaminophen with autism. In the cord blood and meconium studies, they divided those levels into three groups or tertials, the lowest third, the middle third, and the highest third. Then they compared the rates of ADHD and autism in children from each of these thirds. Here’s what they found.
[16:25 –> 16:57] In cord blood, children in the middle tertial had about a 1.69 higher risk of ADHD compared to those in the lowest tertial. In the highest tertial, the risk was two point eight eight times. For autism, the middle tertial showed a one point three eight times higher risk, and the highest tertial showed a three point seven two times higher risk. In the meconium studies, being in the exposure group was linked to a two point four three times higher risk of ADHD. Those numbers sound more dramatic because here we’re not relying on self reporting.
[16:57 –> 17:36] We’re actually measuring acetaminophen in the baby’s cord blood and meconium, but even these studies are still subject to the same problems as the prospective ones, confounding factors. It may be that women who had higher acetaminophen levels were also influenced by other factors, genetics, environment, lifestyle, and these things made them both more likely to use acetaminophen during pregnancy and more likely to have children who developed neurodevelopmental disorders. So acetaminophen may not be the cause at all. Instead, both the acetaminophen exposure and the child’s later diagnosis could simply be different branches of the same root cause. That’s the challenge here.
[17:36 –> 17:57] Even when you see what looks like a dose response relationship in cord blood or meconium, you still can’t prove causation. So with those issues, we looked towards the sibling studies. These were the largest of all, involving the most patients. The biggest one was conducted in Sweden and analyzed about two point four to two point five million births. And here’s how it worked.
[17:57 –> 18:34] Researchers looked at what they called discordant siblings, meaning children born to the same mother where acetaminophen was used in one pregnancy but not in the other. That setup is powerful because it controls for so many confounding factors like genetics, home environment, and environmental exposures. The major difference between the siblings is whether or not acetaminophen was used during pregnancy. And when they made that comparison, the association disappeared. There was no increased risk of ADHD or autism spectrum disorder with acetaminophen use once siblings were directly compared.
[18:34 –> 19:17] That’s a strong contradiction to the signal seen in the perspective in cord blood studies. And this is an example of what’s called familial confounding, the idea that some families carry traits that both make acetaminophen use during pregnancy more likely and independently raise the risk of ADHD or autism. If you don’t fully measure or control for these traits, you can end up mistaking a family driven issue for a drug effect. Norway produced a similar study, again, when pregnancies were compared within the same family where one child was exposed in utero to acetaminophen and the other was not, the apparent association with ADHD vanished. Now these sibling studies were included in the review paper, but the authors downplayed their impact.
[19:17 –> 19:54] In the Swedish cohort, for example, medication use was captured through midwife interviews and prescription dispensing records, yet only seven point five percent of women reported taking acetaminophen during pregnancy. That didn’t line up with other data that showed that more than half of the pregnant women in the world used acetaminophen. Review authors argued that this low reporting rate suggested acetaminophen use was being underestimated, and therefore, the null result couldn’t be trusted. In their view, the sibling studies didn’t move the needle against the associations that were seen elsewhere. But here’s one thing that doesn’t line up with that objection.
[19:54 –> 20:32] The Norway study actually didn’t show low acetaminophen use, and it also showed no association of prenatal acetaminophen use with ADHD within families. So it’s worth emphasizing, these were the largest, most rigorous studies in the dataset, And rather than reinforcing an association, they contradicted it outright. On top of the human studies, the authors also included experimental models. Animal cell studies have shown that acetaminophen can alter brain development and behavior through mechanisms like oxidative stress, hormonal signaling, and gene regulation. These findings don’t come from experiments in people.
[20:32 –> 21:03] They’re lab models that provide biological plausibility for the associations being reported between prenatal acetaminophen use and ADHD or autism. That said, there are some important caveats. Animal studies often use very high doses, continuous dosing, or different timing of exposure across gestation. That’s not how acetaminophen use is typically used in humans. In real life, women take it intermittently for a headache or fever, and it would not be taken in the sustained way that these animal models sometimes simulate.
[21:03 –> 21:37] And we’re also talking about different species altogether with different gestational periods and completely different genetics and biology. So when you put it all together, the review considered four categories of evidence. Perspective studies, dose response studies from cord blood and meconium, sibling comparison studies, and mechanistic work in animals and cells. To analyze them, the authors used something called triangulation. Triangulation means combining evidence from different study designs, each with its own strengths and weaknesses to see if they all point in the same direction.
[21:37 –> 22:12] Since observational data can’t prove causation and randomized control trials can’t ethically be done in this case, triangulation is used as a way of strengthening the signal. The idea is that the weakness of one study might be offset by the strengths of another. And using triangulation, what did they conclude? They argued that taken together, the evidence supported an association between prenatal acetaminophen use and increased risk of ADHD and autism. They acknowledged that the sibling studies showed no effect, but dismissed them on the grounds that acetaminophen use may have been underreported.
[22:12 –> 22:47] Since those studies didn’t show a protective effect, the authors essentially treated the null result as still consistent with their hypothesis. From there, they issued a recommendation. Quote, appropriate and immediate steps should be taken to advise pregnant women to limit acetaminophen consumption to protect their offspring’s neurodevelopment, close quote. Okay. So they also noted that analgesic alternatives like NSAIDs, such as ibuprofen and naproxen, carry teratogenic and developmental risks, especially in the third trimester, and current guidance is to not use them after twenty weeks.
[22:47 –> 23:23] At the same time, they recognize that there are no real pharmacological alternatives for treating fever during pregnancy beyond hydration and physical cooling, and that untreated high fever itself can harm the fetus. So in summary, their position was that acetaminophen use in pregnancy should be restricted. But major regulators like the WHO continue to recommend acetaminophen as the safest option, and to be used at the lowest effective dose for the shortest necessary time. And they emphasized that uncontrolled pain or untreated fever during pregnancy carries its own very real risks. So there we have it.
[23:23 –> 23:38] That’s the data from the study, and that’s what president Trump was referencing. So how do we make sense of all this? First, we can’t afford to be dogmatic here. We have to look at the data with a clear objective eye and without fear. So let’s recap what we know.
[23:38 –> 24:17] This review shows an association between prenatal acetaminophen use and ADHD or autism, and it is very important to remember that this is not proving causation. Association doesn’t prove that acetaminophen is the cause, but it also doesn’t mean that it isn’t. And we have to acknowledge that the evidence is purely associative and weak because of the contradiction in the sibling studies. Those sibling studies controlled for the confounding factors present in the prospective and cord blood studies. So even though this paper came out at Mount Sinai and Harvard, it’s still a systematic review of studies from around the world that showed often weak and contradictory data.
[24:17 –> 24:44] The best we can say is that there may be an association between acetaminophen and neurodevelopment disorders, but causation is unknown. So how would I advise my wife if she were pregnant? Well, we’ve had our four children, but if we were going through pregnancy again, here’s how I’d think about it. I’d tell her, let’s treat acetaminophen like any other chemical input during pregnancy. If you don’t need it, why take it?
[24:44 –> 25:06] All medications are, at their core, are toxins given at doses below the toxic threshold. So if you can avoid them, avoid them, especially in pregnancy. But sometimes life isn’t that simple. If you’re pregnant and suffering with severe pain or headaches, or if you have a fever, should you just push through and deny yourself treatment because of this new analysis? My answer would be no.
[25:06 –> 25:38] Untreated fever and pain can harm the fetus and cause unnecessary suffering in the mother. Based on this associative evidence, I would tell my wife, if you need it, take acetaminophen, but take it sparingly and at the lowest effective dose. That feels very reasonable and cautious given the current level of data and evidence we have. Now, you might take a stricter view and decide you’ll never touch acetaminophen during pregnancy, and that’s very reasonable. Or you might say it’s been considered safe for decades and you’ll keep using it as needed, and that’s also reasonable.
[25:39 –> 26:11] The key is that you’re making your decisions with the data in hand and not based on fear, politics, or sound bites from people who don’t understand science. That’s been my goal here, to put the evidence in front of you so you can think critically, make informed choices, and stay in control of your own health and your baby’s health. Because knowledge is power. When you rely on headlines or political motives to make your decisions, you give that power away. So I hope this helped clarify the issue and gave you a foundation for your own decision making.
[26:11 –> 26:27] If you strongly disagree with me, that’s fine. That’s the beauty of patient autonomy and informed consent. We all get to make our own decisions about our own bodies. So next week, we’ll shift gears to caffeine. I had already started preparing that episode, and honestly, I was worried.
[26:27 –> 26:48] You know how much I love coffee, and I was bracing myself for bad news, that maybe the cost was not worth the benefit. But spoiler alert, that’s not been my conclusion. Of course, you may hear it and decide otherwise, and that’s fine too. But my job is to put the evidence out there so you can make your own decisions. So until next time, stay curious, stay skeptical, and stay healthy.
[26:48 –> 26:48] Cheers.