Advances in obstetrical and neonatal medicine have improved the survival of premature infants, especially those born at younger gestational ages. Infants born before 32 weeks gestation now represent more than 2% of all live births and their survival rates exceed 85%.1 Follow-up studies have revealed high rates of neurodevelopmental disability among infants born very preterm who survive.2 Identifying infants who are most at risk for subsequent neurodevelopmental disability and who may benefit from early intervention services is increasingly important. There is evidence that early detection and early intervention can help to ameliorate disabilities resulting from cerebral injury.3
Magnetic resonance imaging (MRI) prior to hospital discharge may provide early detection of cerebral injury and allow families of neonates at high risk for developmental delay to seek early intervention. MRI studies have revealed that many infants born very prematurely have white-matter abnormalities, including signal abnormalities, loss of volume, cystic abnormality, enlarged ventricles, thinning of the corpus callosum, and delayed myelination.4–6 Several investigations suggest that lesions detected with advanced quantitative MRI technologies, such as volumetric MRI, before hospital nursery discharge correlate with neurodevelopmental deficits assessed at 1 to 2 years of age.7–9 To date, studies using proton magnetic resonance spectroscopy (MRS) in prediction of neurodevelopmental outcome in infants born prematurely are limited.
We hypothesized that volumetric MRI and proton MRS, conducted prior to discharge from the neonatal intensive care unit (NICU) in infants born prematurely, would aid in predicting motor impairment at 6 months adjusted age.
The study included a cohort of infants born prematurely and recruited for volumetric MRI, evaluation of long-term follow-up, and evaluation of early motor learning using a mobile paradigm. This analysis includes results of the MRI analysis and the first developmental evaluation at 6 months of age. Analyses of data using the mobile paradigm are ongoing and will be reported at a future date. Infants born prematurely at gestational ages less than 30 weeks, free of any significant congenital or chromosomal anomaly, and admitted to the NICU at Christiana Hospital in Newark, Delaware, between November 2007 and December 2009 were eligible. A cohort of infants born at term (gestational ages ≥37 weeks), who were admitted to our NICU with minimal illness, were also recruited as a comparison group. Exclusion criteria for recruited infants born at term included major congenital anomalies, need for mechanical ventilation, seizures, 5 minute Apgar score less than 7, and clinically significant hypoglycemia. Infants were enrolled into the study following informed written parental consent as approved by the Christiana Care Institutional Review Board and the University of Delaware Human Subjects Review Committee.
The best obstetric estimate of gestational age was used to indicate pregnancy length, on the basis of last menstrual period date or first trimester ultrasonographic results. Intrauterine growth restriction was defined as weight of less than 10th percentile for gestational age. Bronchopulmonary dysplasia was defined for those infants requiring supplemental oxygen at postmenstrual age of 36 weeks. For the purposes of this study, infants with radiographic evidence of intestinal dilation, ileus, or pneumatosis intestinalis were defined as having necrotizing enterocolitis. Sepsis was defined as any systemic bacterial infection documented by a positive blood culture. At 1 month of age, all infants less than 32 weeks gestation or with weight less than 1500 g were evaluated for retinopathy of prematurity by pediatric ophthalmologists. Subsequent examinations were determined by clinical findings. Head ultrasounds were done routinely during the first week of life and at 1 month of age. Ultrasounds were graded on the basis of the Papile Classification.10 Severe intraventricular hemorrhage was considered grade III-IV. Infants were considered to have cystic periventricular leukomalacia if they had echolucent cysts in the periventricular white matter.
MRI and 1H-MRS Acquisition
Infants underwent MRI and MRS at 36 weeks postmenstrual age or just prior to hospital discharge if the infants were physiologically ready for discharge prior to 36 weeks postmenstrual age. MRI scans were performed without sedation. Before undergoing MRI, each infant was fed and swaddled, and silicone earplugs were applied. All infants' heart rates and oxygen saturations, as measured by pulse oximetry, were closely monitored during the scan. We performed MRI and MRS using a 1.5-Tesla Siemens Symphony MR unit. For the acquisition of the primary MRI data, a spoiled gradient recalled sequence (1.25-mm axial slices; flip angle, 45°; repetition time, 1730 ms; echo time, 3.49 ms; gap, 50%; field of view, 180 mm; matrix, 256 × 256) was applied. For MR spectroscopy data, a single voxel of 15 mm × 15 mm × 15 mm was centered in the periventricular parietal region and a PRESS sequence using a repetition time of 1500 ms and an echo time of 135 ms was performed. MRI scans were completed at Christiana Hospital and were digitally transferred to a workstation at the University of Delaware, in Newark, Delaware, for volumetric analysis.
Volumetric MRI Analysis
Axial T1-weighted spoiled gradient recalled MRI scans were imported into Amira 5.0 software (Visage Imaging, Inc, Carlsbad, California) for manual segmentation and volume rendering. Prior to segmentation, images were evaluated for usability. Images with excessive motion artifacts were excluded from segmentation and volumetric analysis. Tissue segmentation was performed by manually segmenting low-intensity cerebrospinal fluid ventricular spaces (right and left lateral ventricles, third ventricle, and fourth ventricle) on the basis of pixel intensity and known spatial neuroanatomic boundaries (Figure 1). Subcortical gray matter was manually segmented in a similar manner. Subcortical gray matter was defined as structures with higher signal intensity (consistent with gray matter tissue) medial to the external capsule, lateral to the midline, inferior to the anterior horn of the lateral ventricle, and superior to the third ventricle. These included the thalamus, hypothalamus, globus pallidus, putamen, and claustrum. The MRI scans did not readily differentiate the internal capsule from surrounding gray matter tissue. Therefore, subcortical gray matter volumes also included this white matter structure. Larger structures, including cerebellum, cortical gray matter, and cortical white matter, were identified and labeled last, on the basis of signal intensity and spatial location. The total brain tissue volume was defined as cerebrum plus cerebellum (excluding volume of ventricles). For a given structure, the 2-dimensional segmented area was multiplied by the thickness of the slice, and the resulting 3-dimensional volume was summed for each slice containing the segmented structure, to yield the absolute volume (in mL).
Metabolites of Interest on 1H-MRS
Magnetic resonance spectra are simply graphs of amplitude versus frequency in which the area under a “peak” is proportional to concentration (Figure 2). For in vivo brain tissue, the strongest metabolite signals and the common metabolites detectable at millimolar concentrations by proton MRS are from N-acetylaspartate (NAA), creatine (Cr), and choline (Cho).11N-acetylaspartate is located almost exclusively in neurons and axons. This resonance has been widely regarded as a marker for neuronal function.12 Disorders that reduce neuronal or axonal number and density, impair neuronal function, or cause displacement of neurons demonstrate reduced NAA. Creatine is converted to phosphocreatine, which is a high-energy dependent compound critical for maintaining cellular energy-dependent systems. The overall level of Cr is considered to remain stable; therefore, Cr is used as a standard for comparison with other metabolites.13 Lastly, Cho is the precursor for acetylcholine and phosphatidylcholine, which is a major constituent of the cell membrane. Changes in the level of this metabolite are inherently linked to membrane biochemistry. Choline elevations are regarded as a product of myelin breakdown.12
For objective measurement of illness severity, the score for neonatal acute physiology (SNAP), was used. This score is based on 26 routine clinical tests and vital signs. The SNAP has been previously shown to directly correlate with illness severity and mortality rate and a SNAP higher than 15 correlates well with higher mortality rate.14
Developmental Follow-up at 6 Months
All infants were evaluated in their home environment by a pediatric physical therapist (ML). At 6 months of age adjusted for prematurity, the Bayley Scales of Infant Development, Third Edition (BSID-III),15 was administered by 1 trained therapist. The sessions were video recorded with 2 camera views. A second trained physical therapist blinded to infants' birth and medical histories scored these assessments from video. The BSID-III assesses development (at ages 1 to 42 months) across 5 domains: cognitive, language, motor, social-emotional, and adaptive.15 Scaled scores (mean = 10, SD = 3), composite scores (mean = 100, SD = 15), and percentile ranks for the 3 areas evaluated by item administration (cognitive, language, and motor) are provided. The cognitive composite score is composed of findings from the cognitive scale. The language composite score is composed of findings from the receptive and expressive language scales. The motor composite score is composed of findings from the fine and gross motor scales. A low motor score was defined as performance 1.5 standard deviations below the standardized mean for the BSID-III score.
Sample size for the study cohort was estimated a priori on the basis of analysis of early motor learning as determined by a mobile paradigm (analysis for those data are not presented in this article). Power calculation was based on our previous publications with infants born with very low birth weight, and up to 25% more variability than in our previous work. Assuming 0.80 power and P < .05, our calculations yielded estimates of 17 to 30 per group. For statistical analysis, Pearson's correlation, ANOVA, and Mann-Whitney U tests were performed using Statistica version 7. Statistical significance was set at an alpha level of 0.05.
Forty-five babies were recruited into this study. Among these, 38 were born less than 30 weeks gestation and 7 were born at term. Any infants lost to follow up, missing MRI or MRS data sets, or whose MRI images were unable to be segmented for volumetric analysis due to excessive motion artifact or poor quality were excluded from statistical analysis. Among the infants born preterm, 26 had successful MRI and MRS acquisition and 20 had undergone their 6-month follow-up. When comparing the infants born preterm we included in the study with those excluded due to lack of MRI acquisition or loss to follow-up, no differences were found in birth weight (939 ± 238 vs 869 ± 268 g, P = .37) or gestational age (26.5 ± 1.7 vs 26.2 ± 1.8 weeks, P = .54), respectively. Among the infants born at term, 4 successfully underwent MRI, MRS, and the 6-month follow-up. We excluded 1 infant born at term based on imaging findings of a large choroid plexus cyst, which we felt could serve as a confounding variable.
Table 1 lists the Bayley-III composite and subscale scores for the infants born prematurely and the infants born at term. The cognitive composite score of the infants born at term was significantly higher than the infants born prematurely. No differences were found in the motor and language scores.
The infants born preterm when compared with the infants born at term had smaller white matter volume (146.5 ± 6.9 vs 107.7 ± 17.8 mL, P = .001) and gray matter volume (114.6 ± 22.2 vs 144.1 ± 16 mL, P = .04). They were also found to have smaller cerebellar volumes (16.4 ± 5.2 vs 23 ± 1.9 mL, P = .04), subcortical gray matter volume (17.7 ± 3.8 vs 23.9 ± 6 mL, P =.02), and total brain volume (262.7 ± 27.2 vs 337.5 ± 12.1 mL, P < .001) when compared with infants born at term. No differences were found in ventricular volumes in the group born prematurely (12.4 ± 32 vs 2.7 ± 0.3 mL, P = .61) compared with the infants born at term.
The NAA/Cho ratio was higher in the infants born at term than the infants born preterm (P = .046). No differences were noted in the NAA/Cr or the Cho/Cr ratios (data not shown). In investigating the relationship of the metabolite ratios to brain volumes, we found NAA/Cho to be positively correlated with white matter (R = 0.45, P = .03), gray matter (R = 0.43, P = .04), and cerebellar volumes (R = 0.6, P = .002), and thus, total brain volume (R = 0.63, P = .001).
Among the infants born prematurely, 7 of 20 infants had low motor scores. Table 2 shows the characteristics of the infants born prematurely with low motor scores compared with those born prematurely with normal motor scores. Infants with low motor scores had decreased subcortical gray matter volume compared with infants with normal motor scores (Table 3). No differences were found in total brain, cortical white matter, cortical gray matter, cerebellar or ventricular volumes between these groups.
Infants born preterm with low motor scores had similar metabolite ratios as those infants born preterm with normal motor scores (Table 4). No correlations were found between metabolite ratios and individual brain volumes. In addition, metabolite ratios were not noted to correlate with the Bayley motor scores at 6 months.
The main finding of our study was that among infants born prematurely, a reduction in subcortical gray matter volume was associated with delayed motor performance on the BSID-III at 6 months adjusted age. As in previous studies,6,9,11 our data confirm a smaller global brain volume among infants born prematurely. In our study sample, premature birth was associated with reduction in total brain volume due to a decrease in cerebral cortical gray matter, cerebral cortical white matter, subcortical gray matter, and cerebellar volume. A substantial number of surviving infants born prematurely will experience future neurodevelopmental challenges. Our data suggest that MR volumetry may potentially assist clinicians in predicting, before hospital discharge, infants at risk for low motor scores at 6 months adjusted age.
Our finding of reduced subcortical gray matter in infants born preterm with low motor scores has potential important clinical implications. The control of movement and tone in the human nervous system is highly complex and depends on many structures to be intact and provide the right balance of signals to be operational.16 The subcortical gray matter includes the basal ganglia system, which influences muscle power and tone by its effects on the corticospinal tract. A small proportion of our measurements of the subcortical gray matter also included the internal capsule, a white matter structure, which is important in motor control. The observed reduction in the subcortical gray matter volume may be indicative of changes in the signals received and provided by the basal ganglia system. The basal ganglia play an important role in motor behavior. Extensive evidence indicates a role for the basal ganglia in learning and memory. One prominent hypothesis is that this brain region mediates a form of learning in which stimulus-response associations or habits are acquired in increments. Support for this hypothesis is provided by numerous neurobehavioral studies in different mammalian species.17–23 Examples of stimulus-response learning before 6 months of age include reaching for and acting on objects, or infants changing their spontaneous activity level to demonstrate their needs to others. In terms of motor control, if power and tone are atypical, infants' earliest spontaneous movements like kicking and arm flapping, behaviors that teach infants about control of their bodies and movements may be altered. This could affect development as these behaviors are building blocks for later skilled behaviors such as reaching and walking.
In our study sample, 4 of the 7 infants born prematurely with low motor scores had severe intraventricular hemorrhage and/or cystic periventricular leukomalacia on head ultrasounds whereas only 2 of the 13 who had normal motor performance had cystic periventricular leukomalacia and none had severe intraventricular hemorrhage. Similarly, infants with low motor scores had a high rate of sepsis. Severe intraventricular hemorrhage, cystic periventricular leukomalacia, and sepsis are all known to be clinically associated with motor and cognitive deficits. From our data we cannot determine if these morbidities may have contributed to the observed reduction in subcortical gray matter or had an independent effect on motor performance. Clinical courses and long-term outcomes of infants born preterm are highly variable, making prediction of these outcomes difficult. Future studies with larger number of infants will be needed to determine if morbidities such as severe intraventricular hemorrhage, cystic periventricular leukomalacia, or sepsis are responsible for reduction in subcortical gray matter in infants born preterm and further delineate the role of volumetric MRI in routinely predicting motor outcomes.
In addition to MRI, which provides clinicians with anatomic information, MRS has the ability to quantify regional brain metabolism. In our study sample, we did not find an association between MR spectroscopy in the periventricular parietal white matter and low motor scores. We cannot rule out the possibility that measurements of metabolites in other brain regions would have yielded different results. In our study, the MR spectroscopy voxel was placed in the periventricular parietal area to capture potential loss of white matter secondary to diffuse (or noncystic) white matter damage, which is more commonly found in infants born prematurely. The 3 metabolite ratios measured in our study were NAA/Cr, NAA/Cho, and Cho/Cr. In early brain development, there is a normal decrease in Cho, which likely reflects incorporation of Cho into macromolecules associated with myelination.24 At the same time, NAA, which is primarily stored and synthesized in neurons, increases as the brain develops. Reduction in NAA likely reflects decreased neuronal viability, neuronal function, or neuronal loss.24 In our study, there was a reduced NAA/Cho metabolite ratio in infants born premature that did not correlate with motor outcome. The finding of reduced NAA/Cho in infants born preterm may be representative of some degree of delayed myelination that was not of sufficient magnitude to affect early motor function. Alternatively, as infants born preterm are known to have delayed myelination,16 this finding may also be representative of normal brain development in infants born preterm.
Numerous studies have found decreased NAA/Cr and/or NAA/Cho ratios to be predictive of abnormal neuromotor outcome in newborns born at term with hypoxic-ischemic encephalopathy.25–28 Our findings are consistent with a recent study which found NAA/Cr and/or NAA/Cho useful in predicting outcome of neonates with asphyxia, but not predictive of future development in infants born prematurely.29 As MRS is an emerging technology, with studies to date on infants born preterm looking at varied populations and brain regions, there is presently a paucity of normal data for infants born preterm. Our data are helpful in showing a reduction in NAA/Cho in infants born preterm compared with term controls. The parietal periventricular metabolite ratios of infants born preterm measured in our study were similar to those measured in the basal ganglia by Augustine et al.29 As with measurement of regional brain volumes in infants born preterm, the establishment of normative metabolite ratio values are needed to further aid clinicians before MRS is ordered routinely.
Our study has several other important limitations. We failed to find associations with metabolite ratios and 6-month Bayley scores. Our study sample size was not calculated on the basis of analysis of volumetric MRI or MRS. Small biochemical alterations secondary to prematurity may not have been detected due to the limited power of the sample size. Detection of small differences in brain volume including white matter and ventricular volume in infants with low motor scores was also limited by the sample size of this early cohort. Another notable limitation in our study relates to the manual volumetric estimation of the brain structures. In T1-weighted images from infants less than 4 weeks of age, regions of unmyelinated white matter have signal values lower than gray matter.30 This creates a possibility for some misclassification of structures. In addition, Bayley assessments were performed at 6 months of age and may not correlate with later motor status. Although, the previous version of the Bayley was shown to be reliable and valid, the Bayley III is relatively new and, to date, limited evidence exists supporting its construct and predictive validity. Finally, it should be noted that our study investigated a sample of infants less than 30 weeks gestation who were relatively stable, without a high rate associated morbidities, and may not be generalizable to other populations of infants born preterm including those with greater illness severity or more advanced gestation.
Neurodevelopmental outcome remains difficult to predict in infants born preterm. In our study sample, early volumetric MRI studies identify reduction in regional brain volume in infants born preterm. Infants born preterm with Bayley scores less than 1.5 SD from the mean at 6 months of age had a reduction in subcortical gray matter volume identified prior to hospital discharge. Proton MRS completed at term equivalent age did not correlate with motor performance at 6 months adjusted age in infants born preterm. Our data suggest that, with further study, volumetric MRI may have some role in helping to counsel parents about early motor development. Studies with larger numbers of infants are now required to determine if volumetric MRI would be superior to more traditional methods of counseling including using findings from head ultrasounds or other clinical markers, such as sepsis, in predicting motor outcomes.
1. Horbar JD, Badger GJ, Carpenter JH, et al. Trends in mortality and morbidity for very low birth weight infants. Pediatrics. 2002;110:143–151.
2. Marlow N, Wolke D, Bracewell MA, Samara M. Neurologic and developmental disability at 6 years of age after extremely preterm birth. N Engl J Med. 2005;352:9–19.
3. Blauw-Hosper CH, Hadders-Algra M. A systematic review of the effects of early intervention on motor development. Dev Med Child Neurol. 2005;47(6):421–432.
4. Maalouf EF, Duggan PJ, Rutherford MA, et al. Magnetic resonance imaging of the brain in a cohort of extremely preterm infants. J Pediatr. 1999;135:351–357.
5. Inder TE, Wells SJ, Mogridge NB, Spencer C, Volpe JJ. Defining the nature of the cerebral abnormalities in the premature infant: a qualitative magnetic resonance imaging study. J Pediatr. 2003;143:171–179.
6. Ajayi-Obe M, Saeed N, Cowan FM, Rutherford MA, Edwards AD. Reduced development of cerebral cortex in extremely preterm infants. Lancet. 2000:356:1162–1163.
7. Inder TE, Warfield SK, Wang H, Huppi PS, Volpe JJ. Abnormal cerebral structure is present at term in premature infants. Pediatrics. 2005;115:286–294.
8. Peterson BS, Anderson AW, Ehrenkranz R, et al. Regional brain volumes and their later neurodevelopmental correlates in term and preterm infants. Pediatrics. 2003;111:939–948.
9. Anderson PJ, Howard K, Bear M, et al. Relationship between regional brain volumes at term equivalent age and cognitive functioning at 2 years in preterm children. Pediatr Res. 2004;55:3294A.
10. Papile LA, Burstein J, Burstein R, Koffler H. Incidence and evolution of subependymal and intraventricular hemorrhage: a study of infants with birth weights less than 1,500 gm. J Pediatr. 1978;92(4):529–534.
11. Vigneron DB. Magnetic resonance spectroscopy imaging of human brain development. Neuroimaging Clin N Am. 2006;16:75–85.
12. Moore GJ. Proton magnetic resonance spectroscopy in pediatric neuroradiology. Pediatr Radiol. 1998;28:805–814.
13. Govindaraju V, Young K, Maudsley AA. Proton NMR chemical shifts and coupling constants for brain metabolites. NMR Biomed. 2000;13:129–153.
14. Maiya PP, Nagashree S, Shaik MS. Role of score for neonatal acute physiology (SNAP) in predicting neonatal mortality. Indian J Pediatr. 2001;68:829–834.
15. Bayley N. Bayley Scales of Infant Development. 3rd ed. San Antonio, TX: Pearson; 2006.
16. Volpe JJ. Neurology of the Newborn. 5th ed. Philadelphia: WB Saunders; 2008.
17. Graybiel AM. The basal ganglia and chunking of action repertories. Neurobiol Learn Mem. 1998;70:119–136
18. Jog MS, Kubota Y, Connolly CI, et al. Building neural representations of habits. Science. 1999;286:1745–1749.
19. Kesner RP, Bolland BL, Dakis M. Memory for spatial locations, motor responses, and objects: triple dissociation among the hippocampus, caudate nucleus, and extrastriate visual cortex. Exp Brain Res. 1993;93:462–470.
20. Kimura M. Role of the basal ganglia in behavioral learning. Neurosci Res. 1995;22:353–358.
21. Teng E, Stefanacci L, Squire LR, Zola SM. Contrasting effects on discrimination learning after hippocampal lesions and conjoint hippocampal-caudate lesions in monkeys. J Neurosci. 2000;20:3853–63.
22. Butters N, Salmon D, Heindel WC. Specificity of the memory deficits associated with basal ganglia dysfunction. Rev Neurol. 1994;150:580–587.
23. Heindel WC, Butters N, Salmon DP. Impaired learning of a motor skill in patients with Huntington's disease. Behav Neurosci. 1988;102:141–1150.
24. Shah DK, Anderson PJ, Carlin JB, et al. Reduction in cerebellar volumes in preterm infants: relationship to white matter injury and neurodevelopment at two years of age. Pediatr Res. 2006;60:97–102.
25. Khong PL, Tse C, Wong IY, et al. Diffusion weighted imaging and proton magnetic resonance spectroscopy in perinatal hypoxic-ischemic encephalopathy: association with neuromotor outcome at 18 months of age. J Child Neurol. 2004;19:872–881.
26. Miller SP, Newton N, Ferriero DM, et al. Predictors of 30-month outcome after perinatal depression: role of proton MRS and socioeconomic factors. Pediatr Res. 2002;52:71–77.
27. Barkovich AJ, Baranski K, Vigneron D, et al. Proton MR spectroscopy for the evaluation of brain injury in asphyxiated, term neonates. Am J Neuroradiol. 1999;20:1399–1405.
28. Groenendaal FR, Veenhoven RH, van der Grond J, et al. Cerebral lactate and N-acetyl-aspartate/choline ratios in asphyxiated full-term neonates demonstrated in vivo using proton magnetic resonance spectroscopy. Pediatr Res. 1994;35:148–151.
29. Augustine EM, Spielman Dm, Barnes PD, et al. Can magnetic resonance spectroscopy predict neurodevelopmental outcome in very low birth preterm infants? J Perinatol. 2008;28:611–618.
30. Barkovich AJ, Kjos BO, Jackson DE JR, Norman D. Normal maturation of the neonatal and infant brain: MR imaging at 1.5T. Radiology. 1988;166:173–180.