Acetaminophen in pregnancy linked to ADHD and autism in children, but sibling data muddy the picture

Does taking acetaminophen during pregnancy increase a child's risk of ADHD or autism? The question has circulated through medical literature for over a decade, generating anxiety among pregnant women and cautious statements from health authorities. A new cohort study from Taiwan, published in JAMA Pediatrics, adds an important wrinkle: the association appears to exist, but it may not mean what it seems to mean.

Two analyses, two answers

The researchers analyzed a large Taiwanese cohort, examining whether mothers who were prescribed acetaminophen during pregnancy had children with higher rates of attention-deficit/hyperactivity disorder (ADHD) or autism spectrum disorders (ASDs).

In the full cohort analysis, positive associations were observed. Women who received prenatal acetaminophen prescriptions were more likely to have children later diagnosed with ADHD or ASD. Taken in isolation, this result aligns with previous observational studies that have reported similar correlations.

But the researchers then applied a sibling-matched design, comparing children born to the same mother where one pregnancy involved acetaminophen exposure and another did not. This approach controls for all fixed maternal characteristics, including genetics, chronic health conditions, socioeconomic factors, and health behaviors that remain constant across pregnancies.

In the sibling-matched analysis, the association disappeared.

What the divergence tells us

When a population-level association vanishes in within-family comparisons, the most likely explanation is confounding. Something about the mothers who take acetaminophen during pregnancy, rather than the acetaminophen itself, may be driving the elevated rates of neurodevelopmental disorders in their children.

The researchers went further, conducting sibling bidirectional analyses that revealed substantial divergence in associations depending on the analytical approach used. This degree of inconsistency signals unaddressed sources of bias that prevent firm conclusions from being drawn, even using the sibling design.

The study's authors are direct about the implication: the evidence does not support a definitive causal link between prenatal acetaminophen exposure and childhood ADHD or ASD.

Why this matters for pregnant women

Acetaminophen (known as paracetamol in many countries) remains the primary pain and fever medication recommended during pregnancy. Unlike ibuprofen or aspirin, which carry documented risks during certain trimesters, acetaminophen has long been considered the safest option. Previous observational studies suggesting a neurodevelopmental risk created genuine clinical uncertainty and public worry.

This study does not fully resolve that uncertainty, but it substantially weakens the case for a causal connection. The pattern the data show, an association in broad population analysis that dissolves under tighter controls, is characteristic of confounded observational findings rather than real drug effects.

Limitations of both approaches

The full cohort analysis is vulnerable to confounding by maternal factors that correlate with both acetaminophen use and offspring neurodevelopment. The sibling design, while stronger, has its own limitations: it requires pregnancies with different exposure patterns within the same family, which may introduce selection effects. The bidirectional inconsistencies the authors identified suggest that even the sibling design cannot fully resolve the confounding structure in these data.

The study used prescription records, which may not capture over-the-counter acetaminophen use. If many women took acetaminophen without a prescription, the exposure classification would be imprecise, potentially diluting or distorting the results in either direction.

For now, the most defensible clinical position remains what it was before: acetaminophen, used at recommended doses and for appropriate indications, is the preferred analgesic during pregnancy. The evidence for harm remains associational, and the strongest study designs continue to erode rather than strengthen that association.