Effect of breastfeeding on non-verbal cognitive development at 2-years of age in very preterm infants and the role of clinical and social risk: results from the EPICE cohort

Abstract

Introduction There is a considerable controversy over the nature and the pathways involved in the relation between breastfeeding and cognitive development in very preterm infants. In a European cohort of very preterm infants we investigated the effect of breastfeeding on non-verbal cognitive development at 2 years of corrected age and variation by clinical and social risk. Methods Data were obtained from the EPICE (Effective Perinatal Intensive Care in Europe) population-based prospective cohort of infants born between 22 and 31 completed weeks of gestation, in 19 regions from 11 European countries, in 2011–2012. Of the 6792 infants discharged alive, 4425 children had a follow-up at 2 years of corrected age. For analysis, we included all infants with valid information on breastfeeding initiation, non-verbal cognitive development, and on clinical and social risk (n = 4032). Data on infant, obstetric and maternal characteristics and on neonatal morbidity were retrieved from medical records, using a structured form filled in by clinicians until discharge from the hospital. At 2 years of corrected age, data on child's health and development, breastfeeding and sociodemographic characteristics were collected using a structured questionnaire completed by parents. Non-verbal cognitive ability was measured applying the Parental Report of Children's Abilities (PARCA-R) except for France that used Ages and Stages Questionnaire (ASQ). We defined three risk groups based on perinatal factors: High-risk [born before 28 weeks of gestation or with a severe neonatal morbidity (severe congenital anomaly, bronchopulmonary dysplasia, intraventricular hemorrhage grade III or IV, cystic periventricular leukomalacia, retinopathy of prematurity, necrotizing enterocolitis)] (n = 1394, 34.6%); Medium-risk (birth 28–29 weeks of gestation or having growth restriction, but no morbidities) (n = 1485, 36.8%); Low-risk (birth at 30–31 weeks with no morbidities or growth restriction) (n = 1153, 28.6%). Social high-risk was defined as maternal age Results Overall, 15.8% (n = 638) of the children presented suboptimal non-verbal cognition. Never breastfed children were more likely to have suboptimal non-verbal cognition (Crude RR = 1.40, 95%CI: 1.20–1.64). Never-breastfeeding was associated with a significant increased risk in the clinical high-risk group (Crude RR = 1.52, 95%CI: 1.25–1.85), but not in the medium or low-risk groups (Crude RR = 1.05, 95%CI: 0.76–1.44, Crude RR = 1.22, 95%CI: 0.81–1.85, respectively). The stratified analysis by social risk found a significant effect in both low (Crude RR = 1.29, 95% CI: 1.08–1.54) and high-risk (Crude RR = 1.91, 95% CI: 1.29–2.83) groups. After adjustment for the region of birth, corrected age at completion of the questionnaire (months), migration status, clinical risk, and social risk, never-breastfeeding remained associated with an increased risk for suboptimal non-verbal cognition (RR = 1.27, 95% CI: 1.07–1.50). There was no interaction with clinical and social risk in adjusted models. Conclusions Breastfeeding promotes non-verbal cognitive development of very preterm children, independently of clinical and social risk. It provides evidence for specifically targeted interventions to protect, promote and support breastfeeding among these most vulnerable infants.