Background: The effect of breastfeeding on cognitive development in the child remains controversial. The commonly observed association between breastfeeding and higher intelligence quotient (IQ) scores in children may result from the numerous confounding factors in this complex relationship. The aim of this study was to examine whether breastfeeding is associated with later differences in children's cognitive and neurological development.
Methods: Secondary analysis from a 9-year follow-up study of children of 982 smokers recruited to a randomized controlled trial of anti-smoking education in pregnancy in a maternity hospital between November 1981 and October 1982; and all ex-smokers and a random 25% sample of non-smokers during the same period (total 1853 women). Cognitive and neurological development of children was assessed by psychologists in schools using the British Ability Scales and Quick Neurological Screening Test (QNST). Data on breastfeeding and factors relevant to child development were drawn from interviews with mothers at home. Main outcome measures were IQ and QNST scores.
Results: A total of 1218 (66%) children were followed up. In total, 61.6% of children were breastfed, median duration being 12 weeks. Before adjustment, breastfeeding was significantly associated with higher total, verbal and visual IQ scores in children, and increasing duration was significantly correlated with IQ scores. Total IQ was 5.49 points higher in breastfed children (P< or = 0.001), but became non-significant after adjustment. Total IQ scores were most strongly associated with maternal education, maternal age, parity, housing and chronic childhood illness. Breastfeeding was significantly associated with better QNST scores on bivariate analysis but not after adjustment. QNST scores were most strongly associated with maternal education and the child's sex.
Conclusions: Breastfeeding was not associated with any crude IQ advantage or difference in neurological soft signs in children at 9 years. The greater IQ score associated with breastfeeding is accounted for by confounding, with maternal and socio-economic characteristics particularly important.