Age-stratified fat-soluble vitamin profiles and exploratory analyses of potential links to disease risk and core autistic symptoms among children with ASD: a propensity score matching case-control study
Objective To evaluate serum levels of fat-soluble vitamins (vitamin D3 (VD3), vitamin D2 (VD2), vitamin A (VA), total vitamin D (total VD), and vitamin E (VE)) in children with autism spectrum disorder (ASD), to analyze their correlation with autistic symptoms, and provide a theoretical basis for clinical nutritional intervention in ASD. Methods A propensity score matching (PSM) case-control study was designed, enrolling 50 children diagnosed with ASD at Wuhan Children’s Hospital from January 2023 to December 2025 and 50 typically developing (TD) children undergoing routine physical examinations. With age, sex, BMI and season of blood sampling as covariates, Analysis of covariance (ANCOVA) was used to analyze the differences in fat-soluble vitamin levels between the ASD and TD groups. Conditional logistic regression was used to analyze the association between fat-soluble vitamin levels and the risk of ASD. Benjamini-Hochberg (BH) false discovery rate (FDR) correction was applied for multiple testing of all the vitamin indicators. Pearson correlation analysis was employed to examine the relationship between vitamin levels and autistic symptoms. Results After adjustment, serum levels of VD3 and VE in the ASD group were significantly higher than those in the TD group (p < 0.05). Significant age-by-group interactions were observed for VE, VD3, and total VD. Age-stratified ANCOVA analysis showed that the log-transformed VD3 (log VD3) elevation in ASD versus TD children was significant only in the older subgroup and increased with age, whereas no significant group differences were detected for total VD or VE within either age stratum. Our findings revealed no statistically significant association between VD3 levels and ASD risk or core symptoms. Conclusion Age-dependent subtype-specific differences of fat-soluble vitamins exist in children with ASD, with higher log VD3 only observed in older autistic patients. Unadjusted analyses showed tentative correlations between VD3 and ASD risk as well as repetitive behaviors, which disappeared after FDR correction. Clinically, VD3 and total VD should be assessed separately by age for autistic children. Routine VD screening, with longitudinal monitoring tailored to age, is recommended to guide personalized nutritional management.