Introduction
Children born very preterm [VPT, <32 weeks gestational age [GA]] are born during a period that is critical for brain development. At the time of VPT birth, important dynamic processes such as neuronal migration, dendritic and axonal arborisation, synaptogenesis, and myelination are ongoing (Kostović and Jovanov-Milosevi, 2006; Lenroot and Giedd, 2006). By term-equivalent age, infants born VPT already display reduced cortical gray matter, white matter, and subcortical volumes and increased cerebral spinal fluid volume, which have been associated with worse neurodevelopmental abilities (Inder et al., 2005; Monson et al., 2016). Differing trajectories of brain development affecting volumes and cortical thickness in VPT children persist into middle and late childhood (Peterson et al., 2000; Kesler et al., 2004; Mürner-Lavanchy et al., 2014; Zhang et al., 2015; Monson et al., 2016).
In addition to altered brain development, VPT infants are at increased risk of brain injury. Up to 30% of very low birth weight preterm infants experience germinal matrix/intraventricular hemorrhage (GMH/IVH) and white matter injury, that can negatively impact neurodevelopmental outcomes (Bolisetty et al., 2014; Mukerji et al., 2015; Holwerda et al., 2016). Post-hemorrhagic ventricular dilatation (PHVD) remains a clinically important complication of particularly large IVH in VPT infants (Brouwer et al., 2014), and has been associated with smaller total brain, deep gray matter and cerebellar volumes as well as poorer neurodevelopmental outcomes (Messerschmidt et al., 2008; Jary et al., 2012; Brouwer et al., 2016). However, some studies in VPT infants with PHVD have found normal neurodevelopmental abilities in the absence of white matter injury (Brouwer et al., 2012; Reubsaet et al., 2017). Therefore, changes in the cerebellum, deep gray matter and cortical white matter may be important determinants for outcome, which has not been fully explored.
Brain development and outcomes are not well-studied in children born VPT before school-age. In addition, protective factors for the rare cases of normal outcome in high-risk VPT cases with persistent ventriculomegaly following early intraventricular hemorrhage (IVH) are not well-understood. The specific aims of this study were to comprehensively describe the structural brain development of VPT children relative to full-term (FT) children at 4 years of age. First, we describe structural measures including volume, surface area, and cortical thickness within cortical, cerebellar and subcortical gray matter as well as cerebral and cerebellar white matter in these two groups. This is the first such detailed comparison at this age. Second, we compare neurodevelopmental outcomes of VPT and FT children. Third, we describe two remarkable, resilient cases from the cohort of VPT children with persistent ventriculomegaly following IVH, and compare their structural brain measures and developmental outcomes to the larger cohort of children born VPT and FT. In addition to investigating differences between structural brain measures and cognitive abilities between children born VPT and FT children at 4 years of age, we explored the structural determinants of positive outcomes despite severe ventriculomegaly.
Discussion
In our study of 4-year-old children born VPT and FT, differential patterns of brain development were evident. After adjusting for total brain volume, VPT children had increased total gray matter, temporal and occipital gray matter, and reduced pallidum, cortical white matter, and cerebellar white matter compared to FT controls. The children born VPT also demonstrated poorer cognitive outcomes on average, although the scores were in average range. In contrast, the two exceptional cases with persistent ventriculomegaly displayed an alternate neurodevelopmental profile with greater cerebral and cerebellar white matter volumes, yet gray matter volumes within the range of FT children. This structural profile in the cases combined with very good outcomes may suggest a protective phenomenon against abnormal outcome. This is the first study to provide detailed structural brain measures and neurodevelopmental outcomes in children born VPT at 4 years of age, and to show the potential sparing of gray and white matter volumes for normal outcome despite ventriculomegaly in early childhood.
Previous studies in older VPT children have demonstrated widespread differences in cortical and subcortical brain volumes (Peterson et al., 2000; Zhang et al., 2015; Monson et al., 2016), with VPT children typically displaying smaller volumes. Affected structures included the cerebellum, deep gray matter and regions in the frontal, temporal, and parietal lobes (Peterson et al., 2000; Kesler et al., 2004). Monson et al. (2016) reported that VPT infants displayed less growth in absolute cortical and subcortical gray matter volumes and sustained discrepancies in white matter between term-equivalent age and 7 years of age compared to FT children, suggesting that brain development in children born VPT does not catch up to their FT peers during early and late childhood. Our findings of increased gray to white matter ratios in the VPT children, despite having overall reduced gray and white matter volumes compared to FT controls, extend these results. As gray matter steadily decreases throughout childhood (Mills et al., 2016; Vandewouw et al., 2020), increased gray matter in the VPT group is consistent with delayed maturation. Longitudinal studies in VPT children between late childhood and early adolescence have also documented delayed maturation in gray and white matter volumes and slower gains in white matter compared to FT children (Ment et al., 2009; Nagy et al., 2009), although this does not always continue past adolescence (Karolis et al., 2017).
Cortical thickness was greater in the VPT children compared to the FT across lobes. Cortical thinning is a fundamental process of brain development during childhood and reflects synaptic and neural pruning as well as an increase in myelin penetration to lower cortical layers (Sowell et al., 2004). From 7 to 12 years of age, VPT children were found to have continuing cortical thinning in frontal and parietal regions while cortical thinning was largely complete in these regions by 7 years of age in FT controls, reflecting a potential delay in this process in the VPT children (Mürner-Lavanchy et al., 2014). Thus, our study provides evidence of delayed cortical thinning in the VPT children, indicating divergent maturational trajectories of cortical thickness at 4 years of age.
The underlying neurodevelopment of children with PHVD and ventriculomegaly with normal cognitive outcomes remains largely unclear. Our two cases with persistent ventriculomegaly following early IVH were surprisingly resilient with respect to their structural brain measures at 4 years of age. Their cortical gray and white matter volumes were in the ranges of the FT children while white matter volumes of the left cerebellar hemisphere were increased at the ipsilateral side of the largest ventricle. As such, these findings may provide a potential explanation for their intact cognitive abilities.
Post-hemorrhagic ventricular dilatation (PHVD) and non-hemorrhagic ventriculomegaly due to complications of severe IVH or perinatal illness in VPT infants are known to increase the risk for poorer neurodevelopmental outcomes in childhood, such as motor disabilities and cognitive impairments (Ment et al., 1999; Murphy et al., 2002; Pappas et al., 2017). Infants with PHVD who also have parenchymal injury and/or undergo neurosurgical interventions (e.g., ventriculoperitoneal shunt insertion), combined with anesthesia and risk of related complications, often have poor cognitive, motor, attention, and visual perception outcomes (Brouwer et al., 2014; Holwerda et al., 2016; Leijser et al., 2018). Despite what is typical in infants/children with severe PHVD (Tam et al., 2011; Jary et al., 2012; Morita et al., 2015; Brouwer et al., 2016), our two cases did not lose significant brain volume (Jary et al., 2012).
The absence of white matter lesions, relatively stable neonatal clinical courses and gradual progressions of ventricular dilatation without need for intervention likely contributed to their typical outcomes (Futagi et al., 2005). In addition, frontal and temporal white matter regions may have compensated for the ventricular dilatation primarily within the parietal and occipital regions. The size of the ventricles in the other VPT children were significantly larger than the FT children on average, yet not as extreme as two cases, which could be due to cortical volume loss related to prematurity or white matter injury. While these cases are certainly exceptional, our findings strengthen the hypothesis that the presence of additional white matter lesions and changes in cortical, subcortical and cerebellar volumes may be more important determinants of neurodevelopmental outcomes compared to PHVD or ventriculomegaly alone.
There are some limitations to consider within the present study. Although obtaining MRI scans on children at 4 years of age is challenging due to compliance, the sample size of the studied cohort at 4 years of age is relatively small. As such, the results of the present study would be strengthened with a larger sample size. A longitudinal study of the structural brain measures and cognitive outcomes would provide greater insight into neurodevelopmental differences between the children born very preterm and term-born children. Further, information on other factors such as nutrition, therapy, or other treatments could be helpful in understanding potential influences on neurodevelopmental trajectories. Additional cases of PHVD and persistent ventriculomegaly would also be beneficial for understanding the neurodevelopmental outcomes of these children. Furthermore, cortical white matter was not analyzed by lobe using our methodology and thus we could not comment on white matter volumes within the different lobes.
In conclusion, the present study demonstrates for the first time that at 4 years of age, children born VPT present with brain dysmaturation characterized by increased cortical gray matter and reduced white matter volumes after adjusting for total brain volume, and poorer cognitive outcomes in comparison to FT children. Both structural findings suggest delayed maturation, as cortical gray matter decreases and white matter increases with age. Our two cases with isolated persistent ventriculomegaly following early IVH, displayed typical neurodevelopmental outcomes and age-appropriate brain volumes suggesting that sparing of gray and white matter volumes, and absence of associated lesions may be important determinants of outcome in those with PHVD and persistent ventriculomegaly. Long-term outcome studies in large cohorts of preterm infants with persistent ventriculomegaly are needed to support this hypothesis. Taken together, the overall results demonstrate altered neurodevelopmental outcomes in children born very preterm as well as the possibility for normal neurodevelopmental outcomes in children with isolated, slowly progressing PHVD.