Early life multiple exposures and child cognitive function: A multi-centric birth cohort study in six European countries

Publication authors

Jordi Julvez, Mónica López-Vicente, Charline Warembourg, Lea Maitre, Claire Philippat, Kristine B. Gützkow, Monica Guxens, Jorunn Evandt, Sandra Andrusaityte, Miguel Burgaleta, Maribel Casas, Leda Chatzi, Montserrat de Castro, David Donaire-González, Regina Gražulevičienė, Carles Hernandez-Ferrer, Barbara Heude, Rosie Mceachan, Mark Mon-Williams, Mark Nieuwenhuijsen, Oliver Robinson, Amrit K. Sakhi, Nuria Sebastian-Galles, Remy Slama, Jordi Sunyer, Ibon Tamayo-Uria, Cathrine Thomsen, Jose Urquiza, Marina Vafeiadi, John Wright, Xavier Basagaña, Martine Vrijheid


Epidemiological studies mostly focus on single environmental exposures. This study aims to systematically assess associations between a wide range of prenatal and childhood environmental exposures and cognition. The study sample included data of 1298 mother-child pairs, children were 6–11 years-old, from six European birth cohorts. We measured 87 exposures during pregnancy and 122 cross-sectionally during childhood, including air pollution, built environment, meteorology, natural spaces, traffic, noise, chemicals and life styles. The measured cognitive domains were fluid intelligence (Raven’s Coloured Progressive Matrices test, CPM), attention (Attention Network Test, ANT) and working memory (N-Back task). We used two statistical approaches to assess associations between exposure and child cognition: the exposome-wide association study (ExWAS) considering each exposure independently, and the deletion-substitution-addition algorithm (DSA) considering all exposures simultaneously to build a final multiexposure model. Based on this multiexposure model that included the exposure variables selected by ExWAS and DSA models, child organic food intake was associated with higher fluid intelligence (CPM) scores (beta = 1.18; 95% CI = 0.50, 1.87) and higher working memory (N-Back) scores (0.23; 0.05, 0.41), and child fast food intake (−1.25; −2.10, −0.40), house crowding (−0.39; −0.62, −0.16), and child environmental tobacco smoke (ETS) (−0.89; −1.42, −0.35), were all associated with lower CPM scores. Indoor PM2.5 exposure was associated with lower N-Back scores (−0.09; −0.16, −0.02). Additional associations in the unexpected direction were found: Higher prenatal mercury levels, maternal alcohol consumption and child higher perfluorooctane sulfonic acid (PFOS) levels were associated with better cognitive performance; and higher green exposure during pregnancy with lower cognitive performance. This first comprehensive and systematic study of many prenatal and childhood environmental risk factors suggests that unfavourable child nutrition, family crowdedness and child indoor air pollution and ETS exposures adversely and cross-sectionally associate with cognitive function. Unexpected associations were also observed and maybe due to confounding and reverse causality.