Acetaminophen in Pregnancy, ADHD, and Autism: What the Sibling Studies Show

What the Federal Announcement Said, and What the FDA Said Underneath It

On September 22, 2025, the Department of Health and Human Services released a report linking prenatal acetaminophen use to autism, and the FDA began the process of changing the drug's safety label. The announcement drew enormous attention, and it is still shaping what patients ask about in the office.

The agency's own language was more careful than the coverage. The FDA's letter to physicians stated that a causal relationship has not been established and described the body of studies as contradictory. The American College of Obstetricians and Gynecologists responded that no reputable study has shown that acetaminophen in any trimester causes neurodevelopmental disorders.

Three studies published since have tested the question with the designs built to answer it. All three point the same way.

What the Sibling-Controlled Studies Found

The problem with asking whether a medication taken in pregnancy causes a condition in the child is that the people who take the medication differ from the people who don't, in ways that are themselves heritable. ADHD runs in families. So do chronic pain, migraine, infection-prone conditions, anxiety, and the tendency to seek treatment. A mother who takes acetaminophen is not a random draw from the population.

Sibling comparison fixes a large part of that. When the analysis compares two children of the same mother, one exposed in utero and one not, everything the siblings share drops out. That includes the mother's genetics, half of the father's, household income, education, neighborhood, and the family's baseline likelihood of a neurodevelopmental diagnosis.

The Swedish study is the largest cohort ever assembled for this question. In conventional analysis it reproduced the familiar signal: ever-use of acetaminophen was associated with ADHD at a hazard ratio of 1.07 (95% CI, 1.05-1.10). Crude ten-year risk was 2.46% in unexposed children and 2.87% in exposed children. Restricting to matched full sibling pairs moved the estimate to 0.98, and the dose-response pattern disappeared. Children with the highest average daily use had a lower point estimate for autism than children with the lowest.

The Negative Control Is the Whole Argument

The Hong Kong study did something that settles the interpretation more cleanly than any sibling comparison. The authors ran a negative control, asking whether acetaminophen dispensed to a woman before she was pregnant predicted her child's later diagnosis.

It did. Prepregnancy exposure was associated with autism at a hazard ratio of 1.12 (95% CI, 1.08-1.17) and with ADHD at 1.24 (95% CI, 1.20-1.28). Both were statistically unambiguous, and the ADHD estimate was larger than anything the conventional during-pregnancy analyses produced.

No pharmacology explains that. A drug a woman took before conceiving cannot act on a fetus that does not exist. When an exposure that could not possibly have caused an outcome nonetheless predicts it, what the association is measuring is the kind of person who takes the drug. Accordingly, the same mechanism explains the during-pregnancy associations in the older literature, and no biological pathway needs to be invoked at all.

The Danish study triangulated the same point from two more directions. Comparing exposed children with children whose mothers redeemed acetaminophen prescriptions before but not during pregnancy gave a risk ratio of 1.01 for ADHD and 1.01 for autism, which the authors read as confounding by underlying maternal illness. A paternal negative-exposure comparison gave an ADHD risk ratio of 1.23, higher than the primary estimate, which points again at unmeasured familial confounding rather than anything transmitted through the pregnancy.

Why the Conventional Studies Keep Finding Something

Consider what sends someone to the medicine cabinet during pregnancy. Fever from infection, migraine, chronic back or pelvic pain, and inflammatory conditions. Each of those is independently associated with neurodevelopmental outcomes in offspring, each is more common in women with psychiatric and chronic pain diagnoses, and each clusters in families.

Maternal ADHD is the clearest case. A woman with ADHD is substantially more likely than a woman without it to have a child with ADHD, for reasons that are mostly genetic. She is also more likely to have chronic pain, migraine, sleep disruption, and injuries. The acetaminophen is a marker of her situation, and her child's ADHD is largely inherited from her. A study that cannot see her diagnosis records that association as a drug effect.

That is the structure of almost every reported acetaminophen finding, and it is why 43.3% of the Hong Kong cohort having filled a prescription matters. When nearly half a population is exposed, exposure is close to a proxy for engagement with medical care.

The Risk Sitting on the Other Side of the Decision

The conversation usually stops at whether acetaminophen is safe, and it should not, because declining to treat is also a choice with consequences.

Acetaminophen is the first-line analgesic and antipyretic recommended in pregnancy worldwide, and the reason is that the alternatives are worse. Nonsteroidal anti-inflammatory drugs carry documented fetal risk in the third trimester, including premature closure of the ductus arteriosus and effects on fetal renal function. Sustained high fever in the first trimester has its own literature on neural tube defects and adverse outcomes. A pregnant woman with a fever of 103 who takes nothing is not in a neutral position.

Untreated pain and untreated fever are not safe defaults. They are exposures too, and the evidence base behind them is older and more consistent than anything in the acetaminophen literature.

What I Tell a Patient Who Is Pregnant and Asks

The clinical answer is short and has not changed with any of this.

• Treat fever. Acetaminophen at standard doses is the recommended agent in pregnancy, and high fever carries documented risk.
• Use the lowest effective dose for the shortest necessary time. This was already standard advice for every medication in pregnancy and it does not depend on the autism question being settled.
• Don't substitute an NSAID on your own. Ibuprofen and naproxen carry third-trimester fetal risks that acetaminophen does not.
• If you already took it, nothing here suggests you harmed your child. Three studies covering more than three million pregnancies, using the designs built to detect exactly this, found no effect.
• If your child has ADHD and you're searching for a cause, the inherited contribution is large and the single best predictor is a parent with the condition. Many parents recognize their own history in the adult presentation while their child is being evaluated.

What Would Change This

The current evidence has limits worth stating plainly. First, the Danish study followed children only to age 6, so it speaks to early-diagnosed cases, which tend to be the more severe ones. Second, sibling designs assume that exposure differences between pregnancies are not themselves caused by something that also differs, which is an assumption and not a fact. Third, registry studies capture dispensing or recorded use and miss some over-the-counter consumption.

Nevertheless, three independent populations on three continents, analyzed with sibling control, prescription-based ascertainment, paternal negative controls, and prepregnancy negative controls, produced the same null. A finding that survives that many designs is not fragile.

What would move the question is a study with biomarker-confirmed exposure, prospective measurement of maternal indication, and follow-up past childhood. Until that exists, the defensible statement is the one the data support. Prenatal acetaminophen has not been shown to cause ADHD or autism, and the associations in the older literature behave exactly as confounding behaves. Clinicians should remain vigilant about the harm that comes from leaving fever and pain untreated in pregnancy.

Frequently Asked Questions

Does Tylenol during pregnancy cause ADHD?

No study using a design that controls for family confounding has shown that it does. The largest, a Swedish cohort of 2,480,797 children, found a sibling-control hazard ratio of 0.98 (95% CI, 0.94-1.02) for ADHD. A Hong Kong sibling-matched study of 708,020 mother-child pairs found 1.01 (95% CI, 0.93-1.08), with no dose-response.

Why did the FDA announce a label change in 2025?

HHS released a report on September 22, 2025 linking prenatal acetaminophen to autism, and the FDA began the process of a safety label change. The FDA's own letter to physicians stated that a causal relationship has not been established and that the studies are contradictory. The American College of Obstetricians and Gynecologists said no reputable study has shown causation.

What is a negative control and why does it matter here?

A negative control tests an exposure that could not plausibly cause the outcome. In the 2026 Hong Kong study, acetaminophen dispensed before pregnancy was associated with the child's autism (HR, 1.12) and ADHD (HR, 1.24). A drug taken before conception cannot affect a fetus, so that association has to be confounding, and the same confounding explains the during-pregnancy results.

Is it safer to take nothing?

No. Acetaminophen is the first-line treatment for fever and pain in pregnancy because the alternatives carry documented fetal risk, and sustained high fever in early pregnancy has its own established risks. Declining treatment is a choice with consequences, not a neutral option.

I took Tylenol while pregnant and my child has ADHD. Did I cause it?

The evidence says no. ADHD is substantially heritable, and the strongest predictor is a parent with the condition. More than three million pregnancies have now been analyzed with sibling and negative-control designs without finding an effect of acetaminophen.

What about NSAIDs like ibuprofen instead?

Nonsteroidal anti-inflammatory drugs carry third-trimester fetal risks, including premature closure of the ductus arteriosus and effects on fetal kidney function. They are not a safer substitute, and switching without obstetric advice is not recommended.

Further Reading

• ADHD and Dementia Risk — How the same confounding problem shapes a very different ADHD headline
• ADHD in Women — Why so many mothers recognize their own symptoms during a child's evaluation
• Is ADHD Genetic? — Twin studies, heritability estimates, and what a family history actually predicts
• ADHD in Pregnancy and Postpartum — Managing your own ADHD across pregnancy, including medication decisions
• Adverse Outcomes of Untreated ADHD — Pillar resource on what the condition costs when it goes untreated