Impact of prenatal antidepressant exposure on trajectories of childhood emotions and behaviors: evidence from a birth cohort
摘要
Inconsistent findings regarding the long-term effects of prenatal antidepressant exposure on child behavior may stem from a lack of attention to exposure propensity.
MethodsWe used data from the Longitudinal Study of Australian Children (LSAC), a nationally representative birth cohort. Prenatal antidepressant exposure was based on self-reported use during pregnancy. Emotional problems and hyperactivity were assessed at ages 4, 6, and 8. We applied inverse probability weighting (IPW) to growth curve models (GCMs) and repeated measures mixed models (RMMMs) to evaluate behavioral trajectories.
ResultsPrenatal antidepressant exposure was not significantly associated with baseline emotional problems or hyperactivity. However, exposed children showed a steeper increase in emotional problems over time (GCM interaction: β = 0.05, p = 0.003; RMMM age 6 vs. 4: β = 0.12, p < 0.001; age 8 vs. 4: β = 0.09, p = 0.006). Hyperactivity differences became significant only at age 8 (GCM interaction: β = 0.15, p < 0.001; RMMM age 8 vs. 4: β = 0.30, p < 0.001).
ConclusionsWhile not associated with immediate behavioral differences, prenatal antidepressant exposure may shape the developmental trajectory of emotional and behavioral outcomes, emphasizing the need for long-term monitoring.
ImpactPrenatal antidepressant exposure is not associated with immediate emotional or behavioral differences in early childhood, but may influence developmental trajectories, with emotional problems progressively increasing and hyperactivity emerging later. This study highlights the value of modeling symptom trajectories over time and incorporates inverse probability weighting to better account for exposure propensity and reduce confounding bias. The findings support the need for long-term monitoring of exposed children and inform clinical and policy efforts to integrate maternal mental health care with early developmental surveillance.