Autism risk rises with multiple medications taken during pregnancy

Association between cholesterol pathway disruption and neurodevelopment

Cholesterol is a fundamental component of cell membranes and is vital for many molecular processes. Its biosynthesis is especially important during fetal development. Early in gestation, sterols are supplied by the mother, but by mid-gestation, the fetal brain begins producing its own cholesterol. Disruptions in this pathway, due to genetic mutations or external factors, can result in developmental and intellectual disorders, including Smith-Lemli-Opitz syndrome (SLOS), lathosterolosis, and desmosterolosis.

SLOS is the most extensively studied of these disorders, characterized by intellectual disability and distinct physical malformations. It stems from DHCR7 mutations that block conversion of 7-dehydrocholesterol (7-DHC) to cholesterol, leading to low cholesterol, excess reactive 7-DHC, and downstream cellular dysfunction.

About 75 % of people with SLOS are also diagnosed with autism spectrum disorder (ASD), and recent studies link cholesterol metabolism abnormalities to ASDs in general. While rising ASD rates are partly due to better diagnostics and broader criteria, environmental factors may also contribute. Maternal antidepressant use during pregnancy, for example, has been debated as a potential risk factor, especially given links between cholesterol balance and depression risk.

Common medications, such as aripiprazole or trazodone, can elevate 7-DHC similar to SLOS without DHCR7 mutations. Several widely used drugs, including haloperidol, sertraline, and fluoxetine, inhibit cholesterol biosynthesis in multiple models. While early research raised concerns about prenatal exposure and developmental risks, there is a significant gap in large-scale studies examining the outcomes of prenatal inhibition of sterol biosynthesis.

Assessing the relationship between SBIM exposure and ASD risk

Data was obtained from Epic Cosmos, a de-identified dataset of 300 million patient records from over 1,880 hospitals and 42,400 clinics, accessed between March 2025 and January 2026. The current study considered all mother-child pairs with births between January 2014 and December 2023, with follow-up to at least 18 months after birth and through available records up to January 2026. Any non-U.S. births, cases with missing sex, and pregnancies exposed to valproic acid were excluded.

The main outcome was ASD, identified using 51 diagnostic codes. This study specifically focused on drugs taken before or during pregnancy and those that block sterol biosynthesis, such as post-lanosterol pathway inhibitors (e.g., aripiprazole) and 3-hydroxy-3-methylglutaryl-coenzyme A (HMG-CoA) reductase inhibitors (e.g., lovastatin, atorvastatin, simvastatin, rosuvastatin, and pravastatin).

Patients were grouped by the number of unique SBIMs received during pregnancy. Comparator drugs, commonly prescribed in pregnancy but without sterol biosynthesis inhibition, included diphenhydramine, docusate, famotidine, ferrous sulfate, ondansetron, and polyethylene glycol, which served as a reference to assess association specificity.

Sex and preterm birth (<37 weeks) were considered but not modeled, as sex is associated with ASD but not exposure, while preterm birth may lie on the causal pathway rather than act as a confounder. Covariates included maternal age, diabetes, pre-eclampsia/eclampsia, race, ethnicity, year of birth, urban/rural residence, social vulnerability index, tobacco/alcohol use in pregnancy, and pre-pregnancy body mass index (BMI).

Cumulative in utero exposure to SBIM is associated with increased ASD risk in children

The current study analyzed a nationwide cohort of over 6 million US-born children with at least 18 months of follow-up. Of these, 11 % of mothers were prescribed at least one medication affecting sterol biosynthesis during pregnancy, and this rate increased over the study period.

However, several maternal characteristics, including higher rates of metabolic and psychiatric conditions, differed between exposure groups, and these underlying conditions may also contribute to ASD risk.

The association between SBIM exposure and ASD diagnosis rose steadily with each additional SBIM prescribed during pregnancy, reaching over twice the baseline risk when four SBIMs were used. This pattern suggests a cumulative effect from multiple exposures. In contrast, medications that do not affect sterol biosynthesis showed only minimal increases in ASD risk, which further decreased after sensitivity analyses. This supports the specificity of the association, though it does not establish causality.

Sensitivity analyses that accounted for maternal psychiatric diagnoses modestly reduced several associations, particularly for psychotropic medications, but did not eliminate the overall signal.

Conclusions

The authors recommend that clinicians reevaluate their prescribing of SBIMs and closely monitor drugs with sterol-inhibiting activity, including new medications. They also advise educating providers and patients, seeking safer alternatives, limiting multiple SBIM use during pregnancy, considering genetic risk, and supporting additional research. They further caution against overinterpreting the findings and note that many of these medications remain essential treatments, particularly outside of pregnancy, underscoring the need to balance potential risks with clinical benefits.

Download your PDF copy by clicking here.

Journal reference:

• Peeples, E. S., Anzalone, A. J., Dai, R., Reisher, E., Korade, Z., & Mirnics, K. (2026). Sterol pathway disruption in pregnancy: A link to autism. Molecular Psychiatry. 1-11. DOI: https://doi.org/10.1038/s41380-026-03610-7. https://www.nature.com/articles/s41380-026-03610-7