 |
|
 |
|
 |
 |
|
 |
 |
 |
 |
 |
 |
 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
* Division of Psychological Medicine, Institute of Psychiatry, London, United Kingdom
Department of Paediatrics, University College Hospital, London, United Kingdom
 |
ABSTRACT |
|---|
 TOP ABSTRACT METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Methods. Unmatched case-control study was conducted of 33 individuals from a cohort of VLBW (<1500g) infants who were born between 1966 and 1977 and 18 of their normal birth weight siblings. Whole brain, gray matter, ventricular, corpus callosum, and hippocampal volumes were measured on structural magnetic resonance imaging scans.
Results. VLBW individuals had a 46% increase in total ventricular volume and a 17% reduction in posterior corpus callosum volume. No differences in whole brain, gray matter, or hippocampal volumes were observed.
Conclusion. Specific differences exist in the volumes of certain brain structures in adults who were born VLBW compared with their normal birth weight siblings.
Key Words: very low birth weight • brain • volume • MRI • adult
Abbreviations: VLBW, very low birth weight • VPT, very preterm • PVL, periventricular leukomalacia • MRI, magnetic resonance imaging • CSF, cerebrospinal fluid
Very low birth weight (VLBW; <1500 g) and very preterm (VPT; <33 weeks’ gestation) infants are at increased risk for brain injury in the perinatal period1,2 and consequently of impaired cognitive,3 neurologic,4 and behavioral function5 in childhood. These injuries include germinal matrix and intraventricular hemorrhages, infarction of white and gray matter of the cerebral cortex, and diffuse white matter damage (periventricular leukomalacia [PVL]).6,7 A recent qualitative magnetic resonance imaging (MRI) study on a VPT cohort showed that 56% of brain scans of a group of 15-year-old adolescents who had been born VPT were rated as abnormal by neuroradiologists, compared with only 5% of control subjects.2 These abnormalities included ventricular dilation, thinning of the corpus callosum, white matter deficits, and intraparenchymal cysts. More recently, Nosarti et al8 found decrements in whole brain, cortical gray matter, and bilateral hippocampal volumes and enlargement of the lateral ventricles in a quantitative MRI study of VPT adolescents.
However, no study has followed VLBW infants to examine whether any brain damage consequent on such early insults persists into adulthood. Therefore, we measured whole brain, gray matter, ventricular, and bilateral hippocampal volumes as well as corpus callosum volume in a group of adults who had been born VLBW and their normal birth weight siblings. Given previous findings, we hypothesized that VLBW adults would show decreased whole brain volume and corpus callosum volumes and increased ventricular volume compared with their siblings.
|
METHODS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Structural MRI Acquisition
MRI scans were collected on a 1.5T GE Signa system (General Electric, Milwaukee, WI) at the Institute of Psychiatry, London. A series of 
120 slices were formatted in the coronal plane for volume rendering, high resolution of small regions of interest, and arbitrary 2-dimensional slicing. This allowed for good visualization of intracerebral structures in 3 planes. A repetition time of 350 ms and effective echo time of 5 ms were used. A matrix size of 256 x 256 and a field-of-view of 20 cm were set.
Structural brain measurements were rated using the image analysis software MEASURE.9 This Windows-based software allows simultaneous viewing of images in 3 mutually orthogonal planes and uses stereologic principles for volume estimation. The image analysis method used has been described in detail previously.10 Ratings were performed after the entire sample had been collected. Images were mixed and identified only by number so that the investigators were blind to group affiliation. Head tilt was corrected in all brains before any measurements by aligning the brain along the anterior-posterior commissure line in the sagittal plane and along the intrahemispheric fissure in the coronal and axial planes.
A grid size of 5 x 5 x 5 was used to measure the volumes of the whole brain and the gray matter, meaning that every fifth pixel was sampled in the coronal, sagittal, and axial planes. A grid size of 3 x 3 x 1 was used for measuring hippocampal volumes and lateral ventricles, meaning that every third pixel was sampled in the sagittal and axial planes and that every pixel was sampled in the coronal plane. A grid size of 2 x 2 x 1 was used for corpus callosum measurement. All volumes except the corpus callosum were measured by 2 raters (P.F. and P.O.). Corpus callosum volumes were measured by 1 rater (P.A.). The window settings (eg, contrast, brightness) on both computers used for analysis were standardized before interrater reliability testing and remained constant throughout the study. For all volumes, intraclass correlation coefficients for inter- and intrarater reliability, performed on 10 randomly selected independent ratings, all were >0.80.
Morphometric Analysis
Whole brain volume included cortical and subcortical gray matter, white matter, and the brainstem superior to the lower border of the pons. Cortical gray matter volume included the gray matter of the frontal, temporal, parietal, and occipital lobes. The measurement of the total ventricular volume comprised the entire ventricular system, including the lateral, third, and fourth ventricles.
Corpus callosum volume was measured using primarily the sagittal plane for medial slices and a combination of all 3 views to define its lateral, superior, and inferior boundaries. The anterior and posterior boundaries were defined by a line subtended between the most anterior and posterior tip of each lateral ventricle and the intracerebral fissure. Similarly, the superior boundary was defined by a line crossing from the outermost tip of the lateral ventricle to the intracerebral fissure. Inferiorly, the corpus callosum was measured until the lateral ventricle was no longer in view. The corpus callosum was divided into anterior, middle, and posterior portions by 2 lines: one subtended perpendicular to the lowermost posterior point of the rostrum and another subtended perpendicular to the point where the lowermost posterior part of the fornix connects to the body of the corpus callosum.
The volumes of the left and right hippocampus were measured independently, but the rater was not blind to side. In the coronal dimension, the hippocampus was measured rostro-caudally from the first slice where the mamillary bodies were present to the last slice where the fornix was clearly visible. The boundary was defined laterally by the temporal horn of the lateral ventricle; inferiorly by the white matter of the parahippocampal gyrus; superiorly by the alveus; mesially from the temporal horn of the lateral ventricle, along the natural demarcation of the gray matter inferiorly and around to the mesial edge of the temporal lobe; and anteriorly by the beginning of the amygdala.
Ethics
Ethical permission was obtained from the appropriate committees of University College Hospital and the Institute of Psychiatry. Informed, written consent was obtained from each participant before MRI scanning.
Statistics
We used SPSS 1011 for all analyses. We used the 
2 test to compare the distribution of gender and social class between cases and control subjects and the independent-samples t test to compare the mean age of the 2 groups. In addition, the distribution of social class and mean birth weight were compared between the entire original cohort and our study sample. We used 1-way analysis of variance to explore the relationship between whole brain volume between the 2 study groups, controlling for gender and social class. We then used general linear model multivariate analyses with group (case/control status) and gender as factors and whole brain volume, age, and social class as covariates to compare gray matter, corpus callosum, ventricular, and hippocampal volumes between groups. When possible, 95% confidence intervals were calculated, and 2-tailed P values were used throughout the analysis.
|
RESULTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Although whole brain, gray matter, and hippocampal volumes all were reduced in VLBW cases compared with their normal birth weight control siblings, these differences did not attain statistical significance. These findings are in partial agreement with those of Isaacs et al.12 They examined healthy adolescents (n = 11) who were born both preterm and with VLBW and found that their hippocampal volumes were reduced bilaterally compared with full-term adolescents (n = 8) without any evidence of a reduction in whole brain volume.
Ventricular enlargement is a well-recognized complication of premature birth and VLBW. It has been demonstrated both on cranial ultrasound of newly born VLBW infants13 and in adolescents.2 This may result from impaired cerebrospinal fluid (CSF) flow or absorption or from diffuse white matter damage resulting in loss of white matter volume.14 Individuals who experience intraventricular hemorrhage are at increased risk for blockage of CSF flow and consequent pressure hydrocephalus. White matter lesions such as PVL cause loss of tissue, leaving a "gap" that is then filled by CSF (hydrocephalus "ex vacuo"). Our finding of posterior corpus callosum volume reduction might also be explained by such a mechanism. Germinal matrix/intraventricular hemorrhage is the most common brain injury experienced by VLBW and VPT infants and is associated with PVL in most cases.15 Approximately 25% of cases of PVL become hemorrhagic and have a predilection for periventricular arterial border zones, particularly near the trigone of the lateral ventricles. The posterior portion of the corpus callosum lies adjacent to this area and thus represents the part of the corpus callosum most likely to be affected by ischemic changes. Thus, our findings of enlarged ventricles and reduced posterior corpus callosum may well represent the consequences of a single intracranial insult.
Several MRI studies have measured the normal growth of the corpus callosum from childhood to early adulthood using either volume or area measurement. They have consistently reported that the most significant change in the corpus callosum during late development is the increase in its posterior segment (splenium).16–18 Our findings suggest that late developmental changes are unable to compensate for the early insult in the posterior corpus callosum, and because growth in this area may be linked to maturation of the occipital and inferior temporal cortices,17 it is possible that subtle abnormalities may exist in these brain regions in our sample.
A number of limitations of the study merit discussion. Our sample was modest in size, and, given the small volume differences between the 2 groups, it may have had insufficient power to detect volume reductions in all of the regions examined. Indeed, Table 2 demonstrates that all volume measurements in cases were smaller than in control subjects (except for ventricular volume), suggesting that a larger sample size might have been able to detect such differences.
Our cases were chosen 30 years ago on the basis of being VLBW. Recent opinion places more importance on the degree of prematurity than on birth weight in assessing the risk of intracranial lesions. Our sample of VLBW individuals consisted of subjects whose gestational age at birth ranged from 26 to 37 weeks and thus contained a mixture of VPT and almost-term births. The combination of a modest sample size and that our sample was more heterogeneous with respect to size for gestational age may in part explain why we found more modest volume reductions in whole brain, hippocampal, and gray matter volumes than Nosarti et al8 found in their VPT sample. Another explanation may rest with the fact that this cohort was born in an era just before and during the introduction of measures that revolutionized neonatal care, such as incubators and more advanced parenteral feeding techniques. Thus, it may be that those with more severe brain abnormalities may have been less likely to survive than those who were born a decade later, such as those in the later cohort of Nosarti et al.
Finally, we relied on verbal reports from cases and their mothers when recruiting normal birth weight siblings and did not have documentary proof of their exact birth weights. However, our experience when meeting and talking to mothers of cases was that they were usually sure of the birth weight of all of their offspring, probably as a result of having delivered a VLBW infant. Furthermore, there is solid evidence that maternal recall of birth weight is broadly accurate.19–22 Once again, if any of our control subjects were of less than normal birth weight, then the effect would be to minimize the differences between the 2 groups. Thus, although a number of methodologic limitations may be present in this study, their overall effect would be to minimize the strength of our findings rather than to distort them. Consequently, our findings may underestimate the true extent of volume changes in adults born VLBW.
The functional significance of the brain structural abnormalities that we have identified remains unclear. A number of cognitive differences have been described in VLBW compared with normal birth weight individuals. These include lower scores on general intelligence tests in childhood23,24 and adolescence25,26 and in more specific tests of perceptual processing, memory, learning, and problem solving.27 There is also some evidence that VLBW survivors may be disadvantaged in terms of their professional attainment in adulthood compared with normal birth weight adults.25,28 At least 4 follow-up cohort studies have found an association between neonatal cranial ultrasound abnormalities and cognitive function in VLBW children who were not neurologically disabled.29–32 Unfortunately, little is known about the nature of brain structure-function relationships in VLBW survivors beyond childhood. One small study of preterm VLBW adolescents found that bilateral volume reduction in the hippocampus was associated with memory impairment.12 A larger study,33 this time of VPT adolescents, found that cerebellar volume reduction was associated with poorer cognitive performance. A functional MRI study of VPT adolescents34 showed that damage to the corpus callosum is associated with aberrant patterns of cortical activation to visual and auditory tasks and that this may have functional consequences. Our study provides solid evidence that the structural brain consequences of VLBW persist into adulthood and raises the important question of whether these brain abnormalities impair psychological and social functioning in adults who were born VLBW.
|
ACKNOWLEDGMENTS |
|---|
Contribution of authors: Drs Fearon, O’Connell, and Aquino analyzed the MRI scans. Dr Frangou provided training and expertise on MRI analysis and was involved in the statistical analysis, and both Dr Frangou and Dr Allin were involved in revising the manuscript. Drs Nosarti, Taylor, and Rifkin performed the original data collection. Dr Stewart created this cohort and advised on the hypotheses. Dr Murray was involved at every stage of development of this project and was involved in manuscript development. Dr Fearon developed the hypotheses, performed the statistical analyses, and wrote the first and subsequent drafts; he acts as guarantor for the manuscript.
|
FOOTNOTES |
|---|
Reprint requests to (P.F.) Section of General Psychiatry, Division of Psychological Medicine, Box 63, Institute of Psychiatry, London SE5 8AF, United Kingdom. E-mail: p.fearon{at}iop.kcl.ac.uk
|
REFERENCES |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  C. C. Hoppe, R. G. Evans, K. M. Moritz, L. A. Cullen-McEwen, S. M. Fitzgerald, J. Dowling, and J. F. Bertram Combined prenatal and postnatal protein restriction influences adult kidney structure, function, and arterial pressure Am J Physiol Regulatory Integrative Comp Physiol, January 1, 2007; 292(1): R462 - R469. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. desRobert, R. Lane, N. Li, and J. Neu Neonatal Nutrition and Consequences on Adult Health NeoReviews, May 1, 2005; 6(5): e211 - e219. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
||
|
||
Home | My Pediatrics | Journal Information | Current Issue | Past Issues | Subscriptions & Services | Contact Us Click here for faster international access © 2004 American Academy of Pediatrics. All rights reserved. |
||
|
||
