Prenatal Maternal Stress and the Development of Autistic-like Traits in Children: The Iowa Flood Study

• In Event: Poster Session 14
  In Poster Session: PS 14 Topic - Developmental Psychopathology

Abstract

Background: For the past twenty years, the incidence of Autism Spectrum Disorder (ASD) has increased almost five-fold in Canada and the United States. Prenatal maternal stress (PNMS) has been linked to the risk for ASD. For example, research from Project Ice Storm in Canada suggests that the greater the maternal objective hardship and subjective distress from the natural disaster, the more severe the children’s autistic-like traits at age 6 years (Walder, 2014).
However, this approach to studying PNMS and its association with autistic-like traits in children has not been replicated with other disasters, nor has the effect of PNMS on the longitudinal trajectories across early childhood been assessed.
Objectives: Our goal was to determine the effect of prenatal maternal stress from a different disaster on the trajectory of children's autistic-like traits between the ages of 4 and 7, controlling for child gender; although boys appear to be more affected by autism than girls, our previous research in Project Ice Storm found no main effect or gender-by-PNMS interaction.
Method: Beginning one month after the June 2008 Iowa floods, we recruited women who had been pregnant during the floods. We assessed their levels of objective hardship from the floods (threat, loss, scope, change) using the IF100 scale (Yong Ping, 2015). Three scales were used to develop a Composite Scale of Maternal Subjective Stress (COSMOSS): PTSD symptoms as measured by the Impact of Event Scale – Revised (IES-R; Weiss & Marmar, 1997), and peritraumatic distress and dissociation as measured by the PDI (Brunet, 2001) and PDEQ (Marmar, 1996). COSMOSS has a mean of 0 and SD of 1 (Kroska, 2017). Mothers completed the Social Communication Scale (SCQ; Rutter, 2003) to assess their children’s autistic-like traits at ages 4, 5½, and 7 years. A score of 11 on the SCQ is used for screening cutoff for possible autism. We used longitudinal multilevel modeling to test our model with a sample of 217 mother-child dyads.
Results: According to the intra-class correlation coefficient, 53% of the total variance in autistic-like traits is due to inter-individual differences. Age explains 7.5% of the intra-individual variation. As illustrated in Figure 1, the final model shows an effect of COSMOSS subjective distress when adjusted for IF100 objective hardship: for every point increase in subjective distress (COSMOSS) autistic-like traits (SCQ) increased by 0.71 points (p=0.011), while for every additional point on objective hardship (IF100) SCQ increased by 0.41 points (p=0.159). The effect of PNMS on the rate of change was not significant. There was no significant effect of gender on initial status at age 4 years, nor on the rate of change between 4 and 7 years. The significant results remained significant following adjustment for several control variables.
Conclusion: The greater the mother's flood-related subjective flood distress the greater the child’s autistic-like trait score at age 4 years; prenatal maternal distress does not, however, influence the rate of change over time. These results remain significant when adjusted for the effect of the mothers’ objective hardship, and for the child’s age and sex.

Authors

• Mylène Marie-Lyne Lapierre, Université de Montréal (UdeM)
  Presenting Author

• Guillaume Elgbeili, Douglas Research Center
  Non-Presenting Author

• David Laplante, Lady Davis Institute - Jewish General Hospital
  Non-Presenting Author

• Michael W. O'Hara, University of Iowa
  Non-Presenting Author

• Suzanne King, McGill University
  Non-Presenting Author