Maternal and cord blood lipidomics as predictors of autism spectrum disorders: A systematic review.

[en] BACKGROUND: Lipid metabolism is integral to neurodevelopment, contributing to neuronal membrane integrity, myelination, and signaling processes. Recent evidence indicates that disruptions in maternal and perinatal lipidomic profiles may be linked to an increased risk of autism spectrum disorders (ASD). To date, no systematic review has synthesized findings from human cohort studies examining lipidomic biomarkers during pregnancy or at birth in relation to subsequent ASD development.
METHODS: We systematically searched PubMed/MEDLINE, Embase, Scopus, Web of Science, Google Scholar, PsycINFO, CINAHL, and grey literature sources from inception to September 2025 for studies assessing maternal lipidomics during pregnancy, postpartum lipid profiles, or cord/neonatal lipidomics in relation to ASD diagnoses or autistic traits measured in offspring. Eligible study designs included prospective cohorts and nested case-control studies. Data extraction followed a standardized template, and methodological quality was appraised using the Newcastle-Ottawa Scale (NOS). Findings were synthesized narratively given heterogeneity in biospecimen timing, lipidomic platforms, and outcome measures. The protocol was registered with PROSPERO (CRD420251152074).
RESULTS: Nine prospective studies met the inclusion criteria. Maternal lipidomics during pregnancy indicated that lower ω-3 to ω-6 polyunsaturated fatty acid ratios and deficiencies in docosahexaenoic acid were associated with increased autistic traits or ASD with intellectual disability. Postpartum maternal lipid profiles showed that low low-density lipoprotein (LDL) cholesterol predicted greater ASD risk. Cord blood and neonatal lipidomics implicated acylcarnitines, sphingomyelins, and arachidonic acid-derived oxylipins in later ASD symptoms, with some studies demonstrating moderate predictive accuracy (AUROC ranging from 0.71 to 0.85) using machine learning approaches. Overall, recurrent disturbances in fatty acid metabolism, mitochondrial β-oxidation, and inflammatory lipid mediators were observed.
CONCLUSIONS: Prospective evidence supports an association between maternal and neonatal lipidomic alterations and ASD risk, suggesting potential early biomarkers. However, heterogeneity across studies and reliance on single-timepoint measures limit comparability. Standardized lipidomic protocols, longitudinal sampling, and replication in diverse cohorts are needed to establish clinical utility and inform prevention strategies.