The Paradox of Ubiquity: Reassessing the Reported Association Between Prenatal Acetaminophen (Tylenol) Exposure and Autism

Abstract

Background: Observational studies have reported small positive associations between prenatal acetaminophen (Tylenol) use and autism spectrum disorder (ASD). Because acetaminophen is widely used in pregnancy and the indications for its use (e.g., infection, migraine) are themselves associated with neurodevelopmental risk, even tiny imbalances between users and non-users can generate noncausal signals, the paradox of ubiquity.

Objective: To review current human studies on prenatal acetaminophen and ASD, explain how, given the high prevalence of exposure confounding by indication, family factors, and exposure misclassification can distort results, and identify designs that best address these issues, especially sibling analyses.

Methods: We conducted a rapid, structured meta-analysis of human studies through September 5, 2025. Priority was given to: (1) large population-based cohorts with family-based (sibling) comparisons; (2) biomarker-based exposure assessments (umbilical cord blood, meconium); (3) high-quality evidence syntheses (e.g., Navigation Guide); and (4) professional guidance from the American College of Obstetricians and Gynecologists (ACOG) and the Centers for Disease Control and Prevention (CDC). The primary outcome was autism spectrum disorder (ASD); secondary outcomes were attention-deficit/hyperactivity disorder (ADHD) and intellectual disability (ID).

Results: In the nationwide Swedish cohort (n = 2,480,797; 1995–2019), conventional models found marginal risk elevations (e.g., ASD hazard ratio [HR] 1.05, 95% confidence interval [CI] 1.02–1.08), but sibling-control analyses were null (ASD HR 0.98, ADHD HR 0.98, ID HR 1.01), and dose–response patterns vanished; absolute risk differences at age 10 were ≤ 0.21%. Biomarker work in selected cohorts reports higher odds of ASD/ADHD across exposure tertiles, but generalizability and residual confounding remain concerns. Multiple studies link maternal fever, especially in the second trimester and with repeated episodes, to higher ASD risk, consistent with confounding by indication.

Conclusions: With ubiquitous exposures and indication-related use, naïve associations are prone to inflation. The strongest, bias-resistant designs (sibling comparisons) do not support a causal link between acetaminophen and ASD. Clinically, the use of acetaminophen remains appropriate when indicated, at the lowest effective dose and shortest duration. Fever, an independent risk factor, should be treated promptly.