Journal of Developmental Origins of Health and Disease

Original Articles

Associations between birth weight, preeclampsia and cognitive functions in middle-aged adults

P. Factor-Litvaka1 c1, N. Strakaa1, S. Cherkerziana2a3, M. Richardsa4a5, X. Liua6, A. Shera7, G. Neilsa8 and J. Goldsteina2a3a9

a1 Department of Epidemiology, Mailman School of Public Health, Columbia University, New York, NY, USA

a2 Brigham & Women's Hospital Departments of Psychiatry and Medicine, Division of Women's Health, Connors Center for Women's Health & Gender Biology, Boston, MA, USA

a3 Departments of Psychiatry and Medicine, Harvard Medical School, Boston, MA, USA

a4 Medical Research Council (MRC) Unit for Lifelong Health and Aging, London, UK

a5 Department of Epidemiology and Public Health, University College London, London, UK

a6 Department of Biostatistics, Mailman School of Public Health, Columbia University, New York, NY, USA

a7 ICSD.USA, Millburn NJ, USA

a8 Sergievsky Center, College of Physicians and Surgeons, Columbia University, New York, NY, USA

a9 Department of Psychiatry, Division of Psychiatric Neuroscience, Massachusetts General Hospital, Boston, MA, USA


Both reductions in birth weight and preeclampsia (PE) have been associated with decrements in scores on tests of intelligence in children and adolescents. We examined whether these decrements persist into middle adulthood and expand into other domains of cognitive functioning. Using data from the Early Determinants of Adult Health project and from the ancillary project, Fetal Antecedents of Major Depression and Cardiovascular Disease, we selected term same-sex sibling sets or singletons from these sets, from the New England Family Study (NEFS) and the Child Health and Development Studies (CHDS), discordant on either fetal growth or PE, to test the hypotheses that prenatal exposure to inflammation was associated with decrements in attention, learning and executive function 40 years later. Exposure was defined as a continuous measure of percentile birth weight for gestational age, reduced fetal growth (<20th percentile of birth weight for gestational age) or maternal PE. Given that the sample was comprised, in part, of sibling sets, the analyses were performed using mixed models to account for the inter-sibling correlations. Analyses were performed separately by study site (i.e. NEFS and CHDS). We found few statistically significant associations (suggesting a possible type II error) consistent with previous literature, suggesting that the associations with low birth weight do not persist into midlife. We discuss the possible reasons for the lack of associations, which include the possible mediating effects of the postnatal environment.

(Online publication October 14 2011)


c1 Address for correspondence: Dr P. Factor-Litvak, Department of Epidemiology, Mailman School of Public Health, Columbia University, New York, NY, USA. (Email