Prenatal stressful life events predict child cognitive outcomes

Abstract

Introduction: There is growing evidence that prenatal maternal stress is associated with poorer behavioural [1] and cognitive [2] development in the child. What has not been established is whether this is because antenatal stress has a direct effect on the fetus. Animal models indicate that prenatal stress can have a long-term effect on the neuroendocrine stress response in the offspring, specifically the HPA axis [3], although other possible mechanisms include genetic predisposition or postnatal effects mediated by parenting. What is also unclear is the nature of the antenatal stress that is most harmful for child outcomes. We are conducting a study aiming to examine the association between maternal antenatal and postnatal stress and infant cognitive and behavioural development, and to explore possible mediating hormonal mechanisms. Here we present infant outcomes in relation to maternal stress during pregnancy and in the postnatal period. Methods: Women were originally recruited from an amniocentesis clinic, after obtaining written consent. Mothers completed a 26-item questionnaire relating to antenatal and postnatal stressful life events [4]. A trained developmental researcher administered the Bayley Scales of Infant Development (Second Edition) assessing mental (MDI) and motor (PDI) development. The study was approved by the local research ethics committee and is funded by the March of Dimes. Results: Seventy women returned for assessments when their child was about 18 months old ([mean±S.D.] 17.4± 1.1 months). English was the first language in the family and the children (32 girls and 36 boys) were all born at full term (>37 weeks; [mean±S.D.] 39.49±1.26 weeks) and had no clinical problems. Two mothers and babies were excluded because of insufficient English. There was a significant correlation between the total number of stressful life events in the antenatal period and infant MDI [Spearman's correlation (rs)=−0.43, p<0.0001]; the association with PDI was in the same direction but weaker and non-significant (rs =−0.19, n=68). A bivariate analysis showed that the mean±S.D. MDI scores of the children whose mother had experienced three or more life events during pregnancy was 85.8±8.2 (n=15), compared with 98.7±9.0 (n=53) for those children whose mother had experienced under 3 (p<0.0001). In contrast, the association between the total number of postnatal life events and MDI and PDI were non-significant (rs=−.12 and −.02, respectively). In a multivariate analysis including both antenatal and postnatal stressful life event scores, maternal education, antenatal maternal smoking, alcohol and street drug consumption, parity, infant gestational age at birth, infant sex, and infant birthweight for gestational age, the association between antenatal life events and child MDI remained significant (B=−3.19; S.E.= 0.71; β=−0.55; p<0.0001), although birthweight for gestational age also emerged as a significant inverse predictor (B=−2.51; S.E.=2.27; β=−0.12; p=0.05). However, while antenatal life events alone explained 22% of the observed variance in child MDI scores, birthweight for gestational age explained only 6% of the variance. Stressful life events indicative of a strained relationship between the mother and her partner (“you were separated/ divorced”, “you had a serious argument with your partner” and “your partner was emotionally cruel to you”) accounted for 40% of the variance in infant MDI due to life events. Discussion: This study adds evidence to support the hypothesis that stressful life events during pregnancy have a lasting and clinically relevant effect on the cognitive outcome for the child. It supports the concept that intrauterine influences affect behavioural and mental development, independent of postnatal factors. Although birthweight for gestational age emerged as a significant predictor of child MDI scores, the relationship is weak. The data presented here represents half of the final cohort, and future analyses with a fuller cohort will help to clarify this finding. The mediators responsible for the effect of antenatal events are as yet unknown.