Cross-sectional analysis of the association between exclusive breastfeeding duration and Autism Spectrum Disorder odds in U.S. preschool children: a nationally representative study

The association between breastfeeding and Autism Spectrum Disorder remains unclear, with existing evidence showing inconsistency. This study aimed to investigate the relationship between breastfeeding patterns and ASD in a national sample of preschool children. Data from the 2018–2023 National Survey of Children’s Health ( N = 48,158) were analyzed. Breastfeeding was categorized into four mutually exclusive groups based on official NSCH methodology: never breastfed, breastfed < 6 months, breastfed ≥ 6 months non-exclusively, and exclusively breastfed for 6 months. Multivariable weighted logistic regression models examined the association between breastfeeding patterns and ASD, adjusting for confounders. Dose-response was evaluated by modeling breastfeeding category as an ordinal variable (0–3) in fully adjusted models, with polynomial contrasts assessing linearity of trends. Subgroup analyses stratified by age, sex, race/ethnicity, parental education, household size, poverty level, low birth weight, and preterm status tested effect heterogeneity via likelihood ratio tests. Robustness to unmeasured confounding was quantified using E-values for the primary exposure-outcome association. The weighted prevalence of ASD was 3.0%. Compared to never breastfeeding, exclusive breastfeeding for 6 months was associated with significantly lower ASD odds in fully adjusted models (OR = 0.61, 95% CI:0.44–0.86). A significant linear trend was observed, with ASD odds decreasing by 16% for each increase in breastfeeding category (OR = 0.84, 95% CI: 0.76–0.93). Polynomial contrasts further indicated the steepest reduction in odds occurred at the transition to exclusive breastfeeding. Subgroup analyses revealed significant effect modification (P for interaction < 0.05) and stronger inverse associations with lower ASD odds among 3-year-olds, non-Hispanic White children, children of college-educated parents, and those from single-child households. Exclusive breastfeeding for 6 months was associated with lower odds of ASD in U.S. preschool children, with a graded inverse association across increasing breastfeeding exposure categories. This association was modified by sociodemographic and biological factors, underscoring the importance of contextualizing feeding practices within broader family and social environments in neurodevelopmental research. Not applicable.