What just happened, and why people are talking about it
In late September 2025, FDA announced it would start a process to add language to acetaminophen labels noting a possible association with autism and ADHD when used during pregnancy, and it sent a notice to physicians [pdf] summarizing the concern. The move followed statements from the current administration urging caution.
Big caveat: health agencies in the U.S. and abroad quickly emphasized that association is not causation, and that current evidence does not prove acetaminophen causes autism. Newsrooms also highlighted that a federal judge had previously tossed expert testimony in lawsuits claiming prenatal acetaminophen causes autism/ADHD, citing unreliable methods.
Where the medical community stands right now
ACOG (ob-gyns): Reaffirmed that acetaminophen “remains the safest first-line analgesic and antipyretic in pregnancy,” and that the new studies don’t show a clear causal link. Their guidelines are to use the lowest effective dose for the shortest time, as always.
AAP (pediatricians): Fact-checked claims and concluded there’s no causal link between acetaminophen (in pregnancy or childhood) and autism when taken as directed.
Columbia University’s Irving Medical Center (aka Columbia Doctors): Nicely summed up the state of the science as well. Some studies report associations, but larger, more rigorous work, accounting for family factors and genetics, does not support a causal claim.
This is also the stance of other maternal-fetal experts (e.g., SMFM) and international health bodies. Bottom line: treat fever and significant pain in pregnancy—untreated fever carries real risks—using acetaminophen appropriately.
What evidence is the FDA pointing to?
FDA’s communications reference several lines of research. Here’s a plain-English walkthrough.
Narrative reviews and preprints that argue for a risk signal
Prada et al., 2025 (published in: Environmental Health): A Navigation-Guide review concludes there’s “strong evidence” of a relationship between prenatal acetaminophen use and autism/ADHD. Strengths: systematic framing; gathers many observational studies; notes several dose–response signals. Weaknesses: conclusions lean heavily on observational data vulnerable to confounding by indication (why the drug was taken—e.g., fever/infection), exposure misclassification (self-report), and selective emphasis on positive studies. The review’s “strong evidence” is inconsistent with sibling-comparison designs that address family/genetic factors (see Ahlqvist below).
Parker et al., 2023(Children (Basel)): A narrative piece making very strong claims (e.g., that acetaminophen could explain a large share of autism). Strengths: raises mechanistic hypotheses (oxidative stress). Weaknesses: not a randomized trial; relies on ecological and indirect evidence; contains extraordinary claims (even suggesting “most” cases) that are not borne out in higher-quality epidemiologic designs. Treat as hypothesis-generating, not proof.
Recent preprints (e.g., Patel et al.; Yengst’s Florida Medicaid re-analyses): These argue that diagnoses prompting acetaminophen use track with later autism, implying the drug is the driver. Strengths: large datasets. Weaknesses: preprints haven’t undergone peer review; designs are especially sensitive to reverse causation and confounding (children with early neurodevelopmental differences or families with certain health/genetic profiles may experience more fevers/pain, see doctors more, and get more acetaminophen).
Large, rigorous cohort work that tests the confounding problem head-on
Ahlqvist et al., 2024 (JAMA; ~2.5 million Swedish births): In conventional models, acetaminophen use was linked to slightly higher relative risks. But in sibling-comparison analyses, i.e. matching kids within the same family to account for shared genes and environment, the association disappeared for autism, ADHD, and intellectual disability, and there was no dose–response. Why this matters: When a signal vanishes inside families, it points to confounding rather than causation.
This finding lines up with expert summaries from many others. Small associations seen in some studies likely reflect who takes acetaminophen and why, not the drug itself.
The circumcision detour (and acetaminophen by proxy)
A 2015 Danish cohort paper reported higher autism rates among circumcised boys and speculated about pain as a driver. Some have spun this into an acetaminophen story (because acetaminophen is often given after procedures). Problems: the study didn’t measure acetaminophen at all; the significant estimates involved very small numbers; and critics flagged methodological issues (multiple testing, culturally specific patterns, unmeasured confounding). This is not credible evidence that acetaminophen causes autism.
Courts and regulatory decisions about the evidence
In 2023–2024, a federal judge excluded plaintiffs’ causation experts in the acetaminophen–autism/ADHD lawsuits (and later entered judgment[pdf]), pointing to unreliable methods and failure to address confounding. To be clear, court decisions don’t make evidence by themselves, but it mirrors the concerns above.
How should clinicians and families use this information?
-
Do treat fever/pain in pregnancy. Untreated fever has known harms. Acetaminophen, used as directed, remains first-line. Talk with your clinician about dosing and duration.
-
Read headlines with care. “Linked to” or “associated with” ≠ “causes.” When better designs account for family/genetic factors, the acetaminophen–autism link fades out.
-
Look for balance from trusted bodies. ACOG and AAP continue to recommend acetaminophen as needed; Columbia’s experts explain why the causal claim isn’t supported.
A few plain-language takeaways on the science
-
Small associations + big confounding = shaky ground. Many positive studies rely on maternal recall of medication, don’t separate drug effects from the reasons people take the drug (fever, infection, migraine), and can’t fully adjust for genetics or family context. Sibling analyses matter here. And they’re negative.
-
Dose–response isn’t consistent. Some reviews highlight a dose-response effect. The strongest within-family analysis did not find a dose–response with autism.
-
Preprints and narrative essays aren’t the final word. Commentary is useful for developing hypotheses, but not for making policy. Claims that acetaminophen explains “most” autism cases are far beyond what data can support.
Our editorial stance
We aim to be practical and non-partisan. Right now, the best advice hasn’t changed. If you’re pregnant and need to treat pain or fever, acetaminophen is the recommended over-the-counter choice. Use the lowest effective dose for the shortest time, and talk to your clinician. And for everyone reading sweeping claims online: check whether the study design can really tell cause from correlation.
Ben King