Prenatal maternal stress (PNMS) has been linked with adverse health outcomes in the offspring through experimental studies using animal models and epidemiological studies of human populations. The purpose of this review article is to establish a parallel between animal and human studies, while focusing on methodological issues and gaps in knowledge. The review examines the quality of recent evidence for prevailing PNMS theoretical models, namely the biopsychosocial model for adverse pregnancy outcomes and the fetal programming model for chronic diseases. The investigators used PubMed (2000–06) to identify recently published original articles in the English language literature. A total of 103 (60 human and 43 animal) studies were examined.
Most human studies originated from developed countries, thus limiting generalisability to developing nations. Most animal studies were conducted on non-primates, rendering extrapolation of findings to pregnant women less straightforward. PNMS definition and measurement were heterogeneous across studies examining similar research questions, thus precluding the conduct of meta-analyses. In human studies, physical health outcomes were often restricted to birth complications while mental health outcomes included postnatal developmental disorders and psychiatric conditions in children, adolescents and adults. Diverse health outcomes were considered in animal studies, some being useful models for depression, schizophrenia or attention deficit hyperactivity disorder in human populations. The overall evidence is consistent with independent effects of PNMS on perinatal and postnatal outcomes. Intervention studies and large population-based cohort studies combining repeated multi-dimensional and standardised PNMS measurements with biomarkers of stress are needed to further understand PNMS aetiology and pathophysiology in human populations.