ObjectiveStructural brain volumetric differences have been investigated previously in very preterm children. However, children born extremely preterm, at the border of viability, have been studied to a lesser degree. Our group previously analyzed children born extremely preterm at term using voxel-based morphometry. In this study, we aimed to examine regional gray and white matter differences for children born extremely preterm derived from the same cohort during childhood. We also aimed to explore the effect of perinatal risk factors on brain volumes in the same group. MethodsAt 10 years of age, 51 children born extremely preterm (before 27 weeks and 0 days) and 38 term-born controls with high-quality 3.0 Tesla magnetic resonance images were included. Statistical analyses using voxel-based morphometry were conducted on images that were normalized using age-specific templates, modulated, and smoothed. Analyses were also performed in stratified groups of children born extremely preterm in the absence or presence of perinatal risk factors that have previously been shown to be associated with volumetric differences at term. ResultsWe found volumetric decreases in gray and white matter in the temporal lobes, gray matter decreases in the precuneus gyri, and white matter decreases in the anterior cingulum for children born extremely preterm (all p < 0.001, and p(fwe) < 0.05). Gray and white matter increases were predominantly observed in the right posterior cingulum and occipital lobe (all p < 0.001, and p(fwe) < 0.05). Of the examined perinatal risk factors, intraventricular hemorrhage grades I-II compared with no intraventricular hemorrhage and patent ductus arteriosus ligation compared with no treated patent ductus arteriosus or patent ductus arteriosus treated with ibuprofen led to volumetric differences at 10 years of age (all p < 0.001, and p(fwe) < 0.05). ConclusionsChildren born extremely preterm exhibit volumetric alterations in a pattern overlapping that previously found at term, where many regions with differences are the main hubs of higher order networks. Some, but not all, risk factors known to be associated with structural alterations at term were associated with alterations at 10 years of age.
