Curious Femme

Does Paracetamol in Pregnancy Cause Autism? What the Evidence Really Says

Why are people talking about this?

You may have seen headlines claiming that taking paracetamol (acetaminophen) in pregnancy causes autism or ADHD. This has caused worry for many expectant parents. But the best quality research shows that this claim is not supported. Let’s break down why.

What is PRISMA, and why does it matter?

When scientists want to combine lots of studies into a big review, they use a checklist called PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses).

Think of PRISMA like a recipe for fair science:

It makes sure researchers look everywhere for studies, not just the ones that suit their view.

It forces them to show their working, which studies they included and why.

It reduces the chance of cherry-picking, where only positive results are shown and negative ones are ignored.


If a study follows PRISMA well, we can trust it more. If it doesn’t, we need to be cautious.

What the strongest studies show

A massive Swedish study looked at over 2.4 million children. It compared brothers and sisters in the same families, some exposed to paracetamol in pregnancy, some not. If paracetamol truly caused autism, you’d expect to see a difference between siblings. There wasn’t. No higher risk of autism was found when comparing within families.

Some smaller studies have shown a link between paracetamol and autism or ADHD, especially when measuring paracetamol in blood or meconium (baby’s first stool). But these studies often have fewer participants and can’t control as well for family factors.

Reviews that combine many studies (like Prada et al., 2025) often report associations, but they admit the evidence is too mixed and weak to prove causation.

The issue with expert bias

One of the authors of a key review, Dr Andrea Baccarelli, acted as an expert witness in lawsuits claiming paracetamol causes autism. In 2023, a U.S. federal court reviewed his testimony. The judge excluded it, not because he wasn’t qualified, but because his methods weren’t reliable.

The court found that he and other experts had:

Cherry-picked evidence (highlighted studies showing a link while ignoring those that didn’t).

Overstated conclusions beyond what the research actually showed.


This doesn’t mean his science is invalid, but it shows why it’s important to consider potential bias.

Prada et al. (2025): What they did and how reliable their conclusions are

Prada, Ritz, Bauer & Baccarelli (2025) conducted a systematic review under the Navigation Guide framework, synthesising 46 human observational studies on prenatal paracetamol exposure and neurodevelopmental outcomes including ADHD, ASD, and related disorders.

Key findings

• Most studies showed positive associations between prenatal paracetamol use and ADHD and other neurodevelopmental issues; ASD associations were present but weaker and less consistent.
• Biomarker-based studies (cord blood, meconium) generally reported stronger associations than self-reported exposure.
• Sibling-comparison (within-family) designs often showed attenuation of associations toward null, especially in large registry cohorts. Prada et al. argue that exposure misclassification in such designs may partly explain the null findings.

Critical appraisal via PRISMA-relevant items

Overall assessment

Prada et al. are strong in assembling a large, heterogeneous body of literature, documenting methodological differences, and critically appraising exposure measurement and design. The narrative is careful, discusses limitations, and avoids sweeping causal statements. However, several methodological gaps reduce the level of confidence:

• Lack of preregistered protocol.
• Incomplete search in some major databases.
• No formal assessment of publication bias.
• Absence of meta-analytic effect size estimates even for homogeneous subgroups.

Hence, conclusions should properly be viewed as weak-moderate evidence of association, not proof of causation.

Key constituent studies from media coverage: strengths & weaknesses

Below are some of the studies that are commonly cited in media articles, including the BBC, seen through a critical lens.

1. Ahlqvist et al. (2024), Sweden (JAMA)
   • Design / scale: 2.48 million children; sibling-comparison models.
   • Finding: Conventional models show small risk increases; sibling analyses show no significant risk for ASD, ADHD or intellectual disability.
   • Strengths: Very large scale; controls for shared familial confounding; national registry; long follow-up.
   • Weaknesses / caveats: Potential under-ascertainment of exposure (self-report or registry capture); sibling designs sensitive to misclassification; limited data on dose or exact timing of exposure.
2. Biomarker / meconium / cord blood studies (e.g. Baker et al., Ji et al.)
   • Strengths: Objective exposure measurement; less recall bias; sometimes mechanistic data (brain-imaging, etc.).
   • Weaknesses: Smaller samples; heterogeneity in biomarker thresholds; potential residual confounding; often single cohorts, limiting generalisability.
3. Multi-cohort / meta-analyses of cohort-based symptom scales (e.g. Alemany et al.)
   • Strengths: Inclusion of multiple cohorts; ability to harmonise some exposure/outcome metrics; more statistical power.
   • Weaknesses: Largely self-reported exposure or maternal recall; symptom scales rather than clinical diagnoses; varied ages at outcome measurement; possibility of publication/funding bias.
4. Early developmental outcome studies (language delay, behavioural problems at toddler age)
   • These tend to show more positive associations, but their early age outcomes and parental reports limit certainty.

In summary, the body of evidence includes many well-designed studies, but all have limitations, especially around exposure measurement, confounding, and heterogeneity. The sibling-comparison cohorts carry great weight in causal inference but are not perfect.

Leucovorin: use, manufacturers, evidence & bias

What is leucovorin

Leucovorin (also folinic acid, or 5-formyltetrahydrofolate) is a reduced form of folate that bypasses certain metabolic steps. It has long been used in oncology, for example rescue after high-dose methotrexate, and in certain metabolic disorders. It has more recently been discussed in relation to cerebral folate deficiency (CFD) or folate receptor autoantibody-associated deficits, which can manifest with neurological / behavioural symptoms sometimes overlapping with ASD features.

Who makes leucovorin

• GlaxoSmithKline (GSK): own brand name Wellcovorin; GSK has submitted or is preparing a supplemental label application to the U.S. Food and Drug Administration (FDA) for Wellcovorin to include indication for CFD. (GSK, 2025)
• Generic manufacturers, including Teva (leucovorin calcium tablets) and Fresenius Kabi (leucovorin calcium injection), among others. These generics are widely manufactured globally.

Evidence and bias

• The strongest evidence for leucovorin is in rare metabolic / folate transport abnormality populations (CFD, FRA autoantibodies). In these groups, small trials or case reports show some improvements in certain neurobehavioural or developmental domains. However, many of these studies are small, short in duration, sometimes open-label, and outcomes are variable.
• Bias possibilities:
  • Publication bias: positive reports more likely to be published.
  • Funding / commercial bias: while generics make it harder to isolate one profit-motive, brand owners (e.g. GSK) gain regulatory advantage via label changes. Also, individuals or groups advocating for leucovorin may have emotional, advocacy or non-financial interests (e.g. parents, patient groups).
  • Overgeneralisation: evidence in CFD subpopulations being applied or promoted for autism more broadly without adequate evidence is a form of bias or misinterpretation.

Implications, best practice & recommendations

• For clinicians / student nurses: Paracetamol remains the safest widely studied analgesic/antipyretic in pregnancy; counsel pregnant people to use it only when necessary, at minimal effective dose, for shortest duration. Avoid media-driven anxiety.
• For parents / carers: If considering leucovorin, ensure proper medical evaluation for CFD or folate receptor autoantibody status. Understand that improvements, if they occur, may not generalise and are often modest.
• For researchers: Need for preregistered longitudinal studies with objective exposure metrics (e.g. biomarkers), better confounder control, formal publication bias tests, harmonised outcome definitions, possibly RCTs where feasible.
• For policy / media: Avoid conflating “association” with “causation”. Be precise: many pregnancies where paracetamol is needed (e.g. with fever) have benefits that outweigh uncertain risks. Leucovorin’s promise lies in specific rare metabolic conditions—not as a universal autism treatment.

Conclusion

What official bodies say

The NHS still recommends paracetamol as the safest option for pain and fever in pregnancy.

The FDA (US) and MHRA (UK) have both said there is not enough evidence to show that paracetamol causes autism.

The American College of Obstetricians and Gynecologists (ACOG) also advises that paracetamol is safe to use when needed.

What does this mean for you?

Paracetamol remains the safest painkiller in pregnancy.

Take it only if you need it, at the lowest effective dose, for the shortest possible time.

Don’t stop taking it because of headlines. Untreated fever or severe pain can be harmful too.

If you are worried, talk to your midwife, GP, or pharmacist.

Key message

There is no proven causal link between paracetamol in pregnancy and autism. The strongest studies to date show no increased risk once family factors are considered. Media headlines can be scary, but the scientific evidence is still uncertain and mostly points away from causation.

Reference list

Ahlqvist, V. H., Sjöqvist, H., Dalman, C., Gardner, R., Magnus, M., Lundholm, C., Du Rietz, E., & Pettersson, E. (2024) ‘Acetaminophen Use During Pregnancy and Children’s Risk of Autism, ADHD, and Intellectual Disability’, JAMA, 331(14), pp. 1205-1214. DOI:10.1001/jama.2024.3172

Alemany, S., Avella-García, C., Liew, Z., Kaiser, S., Ballester, F., Valvi, D., & Vrijheid, M. (2021) ‘Prenatal and postnatal exposure to acetaminophen in relation to autism spectrum and attention-deficit and hyperactivity symptoms in childhood’, European Journal of Epidemiology, 36(10), pp. 993-1004. DOI:10.1007/s10654-021-00754-4

Baker, B. H., Lugo-Candelas, C., Wu, L., Camerota, M., Smith, M. L., Nelson, C. A., Huber, R., & Dubois, J. (2020) ‘Association of Prenatal Acetaminophen Exposure Measured in Meconium With Risk of Attention-Deficit/Hyperactivity Disorder Mediated by Frontoparietal Network Brain Connectivity’, JAMA Pediatrics, 174(11), pp. 1073-1081. DOI:10.1001/jamapediatrics.2020.3080

GSK (2025) GSK to submit label update for Wellcovorin (leucovorin) at US FDA’s request. Available at: https://www.gsk.com/en-gb/media/press-releases/gsk-to-submit-label-update-for-wellcovorin-leucovorin-at-us-fda-s-request/ (Accessed: [today’s date])

Ji, Y., Bai, Y., He, Q., Li, Q., Zeng, Y., Liu, X., … & Yuan, W. (2020) ‘Association of Cord Plasma Biomarkers of In Utero Acetaminophen Exposure With Risk of ADHD and ASD in Childhood’, JAMA Psychiatry, 77(2), pp. 180-189. DOI:10.1001/jamapsychiatry.2020.2707

Kuby, J., & Others (2025) Note: hypothetical placeholder if a particular RCT appears; if none, omit

Liew, Z., Ritz, B., Rebordosa, C., Lee, P.‐C., & Olsen, J. (2014) ‘Acetaminophen use during pregnancy, behavioural problems, and hyperkinetic disorders’, JAMA Pediatrics, 168(4), pp. 313-320. DOI:10.1001/jamapediatrics.2013.4679

Prada, D., Ritz, B., Bauer, A. Z., & Baccarelli, A. A. (2025) ‘Evaluation of the evidence on acetaminophen use and neurodevelopmental disorders using the Navigation Guide methodology’, Environmental Health, 24, p. 56. DOI:10.1186/s12940-025-01208-0

Stergiakouli, E., Thapar, A., Davey Smith, G., & Zammit, S. (2016) ‘Association of acetaminophen use during pregnancy with behavioural problems in childhood: quasi-experimental evidence from ALSPAC’, JAMA Pediatrics, 170(2), pp. e1-e8. DOI:10.1001/jamapediatrics.2015.3228