RÉSUMÉ
Objective: To investigate the relationship of pregnancy-related anxiety of pregnant women in second/third trimesters and autism-like behaviors in their offspring at 18 months of age. Methods: Based on a prospective cohort study design, we evaluated the situation of pregnancy-related anxiety of women during second and third trimesters through a Pregnancy-Related Anxiety Questionnaire. Subjects under study were classified into three groups, 1) those with pregnancy- related anxiety during both trimesters, 2) those with pregnancy-related anxiety at one trimester and 3) those without pregnancy-related anxiety in either trimester. When their children were 18 months, autism-like behaviors (ALB) were evaluated, using the part A of Checklist for Autism in Toddlers-23, and then classified into three groups as non-ALB group, minor ALB group and major ALB group. Multi-nominal logistic Regression was used to analyze the relationship of pregnancy-related anxiety with autism-like behaviors. Results: Compared with non-ALB group, children whose mother with pregnancy-related anxiety during both trimesters presented significant higher risk on ALB than children whose mother without pregnancy-related anxiety in these two periods (relative risk, RR=2.43, 95% CI: 1.21-4.86, P=0.012), major factors as pregnant women's IQ and gestational diabetes mellitus, premature delivery and education levels of fosterers on these pregnant women were under control. Our results from the stratified analysis showed: when in the subgroup that mother was the main fosterer of the child, there was an significant increase of risk in children whose mothers with pregnancy-related anxiety during both trimesters (RR=4.22, 95%CI: 1.73-10.32, P=0.002). Conclusion: The association between pregnancy-related anxiety and autism-like behavior was not strong but influenced by the fosterer of the child.