“…Factors previously identified from the literature as predictors of postnatal growth (Dubois and Girard, 2006;Regnault et al, 2010;Valvi et al, 2014) or/and chlordecone exposure (Guldner et al, 2010;Seurin et al, 2012) were a priori considered as potential confounders : duration of gestation in weeks (3 categories: 37, 38-39, 40-41), maternal place of birth (French West Indies, other Caribbean islands, Europe), maternal age (continuous), maternal pre-pregnancy height and weight (continuous), maternal weight gain during pregnancy (compliance with pre-pregnancy BMIspecific recommendations, as defined by the US Institute of Medicine (Siega-Riz and Gray, 2013)), parity (0, 1, Z 2), years of education (o 5, 5-12, 4 12), marital status (single, married or in a couple, living with own family), smoking and alcohol consumption during pregnancy (yes/no), duration of breastfeeding (no, o 3 months, 3-7 months, 7-18 months). Missing values for covariates (marital status (N ¼13), alcohol intake (N ¼4), weight gain during pregnancy (N ¼8)) were imputed to the modal category.…”