Moderating effects of maternal-infant interaction on the relationship between adverse neonatal exposures and the development of autistic behaviors at follow-up: A preterm birth cohort study.

BACKGROUND: Preterm-birth infants have elevated risks of autistic behaviors compared to term-birth infants. While neonatal adverse exposures and the morbidity burden in the neonatal intensive care units are often used to forecast neurodevelopmental outcomes including autism, the extent to which post-discharge maternal psychosocial health and mother-infant interactions mitigate early autistic behaviors remains unclear.

AIM: To examine whether the maternal psychosocial status and maternal-infant interaction at 6 months moderate the association between neonatal morbidity exposures and autistic behaviors in a longitudinal preterm cohort.

METHODS: We conducted a cohort study of infants born at <33 weeks' gestation in southern Taiwan. Neonatal morbidity counts (0-5) of severe brain injury, sepsis, necrotizing enterocolitis, severe retinopathy of prematurity, and bronchopulmonary dysplasia were recorded. Family social risks were assessed at discharge. At 6 months of infant's corrected age, mothers completed measures of depression, state anxiety, partnership quality, and the BRIGANCE® Parent-Child Interaction. At 24 months, autistic behaviors were screened using the Modified Checklist for Autism, Revised (>2 as positive), and neurodevelopment was measured by the Bayley Scales of Infant and Toddler Development, Third Edition. Analyses included group comparisons, a multivariable logistic regression, and interaction (moderation) between morbidity counts and parent-child interaction scores.

RESULTS: Of 472 infants who survived to discharge, eight infants died after discharge, and 336 (72.4%) completed the 24-month follow-up. Twenty-six (7.7%) infants screened positive for autistic behaviors. Groups did not differ in gestational age, sex, anthropometrics, or neonatal morbidities. Screen-positive toddlers had lower cognition, language, and motor scores (all p ≤ 0.001). In the adjusted models, lower maternal education increased the odds of autistic behaviors (odds ratio [OR] 4.06, 95% confidence interval [CI] 1.37-12.06), whereas a higher parent-child interaction score was associated with a lower likelihood of autistic behaviors (OR 0.80, 95% CI 0.72-0.89). Higher parent-child interaction attenuated the association between greater neonatal morbidity counts and autistic behaviors (interaction OR 1.31, 95% CI 1.02-1.68). High-quality interactions were associated with lower autistic-behavior prevalence across morbidity strata, especially in those with no neonatal morbidity.

CONCLUSIONS: In preterm infants, post-discharge relational and social factors-especially higher-quality maternal-infant interactions and maternal education-were more strongly linked to autistic outcomes than the neonatal morbidity burden alone. Positive maternal-infant interactions appear to buffer neonatal morbidity vulnerability, supporting the need for routine 6-month screening of the maternal psychosocial status and dyadic interactions, and motivating pragmatic, relationship-focused early interventions to reduce autistic behaviors in high-risk preterm children.