Background. Despite improvements in survival data, the incidence of neurodevelopmental handicaps in preterm infants remains high. To prevent these handicaps, one must understand the pathophysiology behind them. For preterm infants, cerebral ventriculomegaly (VM) may be associated with adverse neurodevelopmental outcome. We hypothesized that although the causes of VM are multiple, the incidence of handicap at 4.5 years of age in preterm infants with this ultrasonographic finding at term would be high.
Methods. To test this hypothesis, we provided neurodevelopmental follow-up for all 440 very low birth weight survivors of the Multicenter Randomized Indomethacin Intraventricular Hemorrhage (IVH) Prevention Trial. A total of 384 children (87%) were evaluated at 54 months' corrected age (CA), and 257 subjects were living in English-speaking, monolingual households and are included in the following data analysis.
Results. Moderate to severe low pressure VM at term was documented in 11 (4%) of the English-speaking, monolingual survivors. High grade IVH and bronchopulmonary dysplasia (BPD) were both risk factors for the development of VM. Of 11 (45%) children with VM, 5 suffered grades 3 to 4 IVH, compared with 2/246 (1%) children without VM who experienced grades 3 to 4 IVH. Similarly, 9/11 (82%) children with VM had BPD, compared with 120/246 (49%) children without VM who had BPD.
Logistic regression analysis was performed using birth weight, gestational age, gender, Apgar score at 5 minutes, BPD, sepsis, moderate to severe VM, periventricular leukomalacia, grade of IVH, and maternal education to predict IQ <70. Although maternal education was an important and independent predictor of adverse cognitive outcome, in this series of very low birth weight prematurely born children, VM was the most important predictor of IQ <70 (OR: 19.0; 95% CI: 4.5, 80.6). Of children with VM, 6/11 (55%) had an IQ <70, compared with 31/246 (13%) of children without VM. Children with VM had significantly lower verbal and performance scores compared with children without VM.
Conclusions. These data suggest that, for preterm neonates, VM at term is a consequence of the vulnerability of the developing brain. Furthermore, its presence is an important and independent predictor of adverse cognitive and motor development at 4.5 years' CA.
- preterm infant
- neurodevelopmental outcome
- intraventricular hemorrhage
- bronchopulmonary dysplasia
- IVH =
- intraventricular hemorrhage •
- VM =
- ventriculomegaly •
- CA =
- corrected age •
- BPD =
- bronchopulmonary dysplasia •
- ECHO =
- echoencephalography •
- PVL =
- periventricular leukomalacia •
- WPPSI-R =
- Wechsler Preschool and Primary Scale of Intelligence-Revised •
- IQ =
- intelligence quotient •
- PND =
- postnatal day
Although the survival statistics for very preterm infants are steadily improving, the incidence of major neurodevelopmental handicaps in very low birth weight preterm infants remains high. Thus, although ∼1% of all live births in the United States weigh <1000 g at birth, and the survival rate for this population of infants is 85%,1 ,2 the incidence of cerebral palsy is increasing because of this increase in survival.3 ,4 Major cognitive handicaps are reported in 5% to 32% of these children.5–10 In addition, at 8 years of age, more than half of these children require special assistance in school, nearly one fifth are educated in designated special education classrooms, and 16% have repeated at least one grade.11–13
If the goals of perinatal intensive care are to improve survival and to prevent handicap in very low birth weight preterm infants, then one must first identify the risk factors for adverse outcome. Thus, during the last 2 decades, numerous investigators have demonstrated that infants with parenchymal involvement of intraventricular hemorrhage (IVH) are at high risk for neurodevelopmental handicap.14–16 In addition, a recent study suggested that cerebral ventriculomegaly (VM) or significant enlargement of the lateral ventricles without evidence of increased intracranial pressure might also represent a sensitive predictor of both cognitive and motor impairment.17
The follow-up phase of a prospective, randomized, placebo-controlled trial of low-dose indomethacin for the prevention or extension of IVH was designed to evaluate the incidence of neurodevelopmental handicap.18 ,19 A total of 505 infants with birth weights from 600 to 1250 g were enrolled; 440 (87%) survived, and 384 (87%) were evaluated at 54 months' corrected age (CA) (ie, age from the obstetric due date). We used this cohort of children to test the hypothesis that moderate to severe cerebral VM at 40 weeks' conceptional age predicts adverse neurodevelopmental outcome at 4.5 years' CA and to determine the risk factors for this ultrasonographic finding.
The clinical studies were conducted at Women and Infants' Hospital, Providence, Rhode Island; Maine Medical Center, Portland, Maine; and Yale New Haven Hospital, New Haven, Connecticut. The protocols and procedures described below were reviewed and approved by the institutional review boards of the three participating institutions.
From September 1, 1989 through August 30, 1992, 505 infants with birth weights of 600 to 1250 g were admitted by 6 hours of age to the participating institutions and, after parental consent was granted, to two parallel randomized prospective trials designed to determine whether the early administration of low-dose indomethacin would either prevent IVH in this patient population (primary outcome) or, in patients with hemorrhage present at 6 postnatal hours, prevent extension of that hemorrhage (secondary outcome).18 ,19 To evaluate the brain for hemorrhage, all infants enrolled in the study were first examined using cranial echoencephalography (ECHO) between 5 and 11 hours. Of the 505 infants who were enrolled, 431 did not have IVH and were randomized to the primary IVH prevention trial. An additional 61 infants with either germinal matrix hemorrhage or small IVH on the initial scan were randomized into the extension trial. All 505 enrolled infants were followed up prospectively and provide the basis of this article.
Subsequent scans to evaluate the brain for hemorrhage and/or ischemic changes were performed at 24 and 48 hours after the first cranial ECHO, on postnatal days (PNDs) 4, 5, 7, 14, and 21, and at 40 weeks' conceptional age or more often if clinically indicated. Scans were interpreted first by the institutional radiologist and later, for data verification, by a central radiologist. In cases of disagreement, the data were reexamined by all participating radiologists, and a group decision was formulated. In each case, radiologic assessment was conducted without previous knowledge of the infant's clinical condition. The grading system for hemorrhages was as previously reported:18 grade 1 indicates blood in the periventricular germinal matrix regions; grade 2, blood within the lateral ventricular system without ventricular dilatation; grade 3, blood within and distending the lateral ventricles; and grade 4, blood within the ventricular system and parenchymal involvement. For the purpose of our analysis, IVH was categorized into three groups: none, grades 1 to 2, and grades 3 to 4. VM was assessed on the ECHO studies performed at 40 weeks' conceptional age (or, if that was unavailable, 21 days of age). Data were available for 234/257 (91%) subjects at 40 weeks' conceptional age; 21-day data were used for the other 23 (9%) children. For the purpose of our analysis, only patients with moderate and severe VM (ie, measurements of 1.0–1.5 and >1.5 cm, respectively, at the midbody of the lateral ventricle on sagittal scan as indicated in Fig 1) were included.20 ,10 The ultrasound studies were also evaluated for the presence of focal echolucencies. All cases identified as showing focal echolucencies had cystic areas consistent with periventricular leukomalacia (PVL) on the ultrasound performed at 40 weeks' conceptional age.21
As previously described, all infants underwent gestational age assessment using a modification of the Ballard scale.22Prenatal, perinatal, and neonatal data were obtained by maternal interviews and prospective review of the maternal and neonatal charts. An infant was diagnosed as suffering from bronchopulmonary dysplasia (BPD) if he/she both required oxygen supplementation and had an abnormal chest radiograph at 28 days of life, as previously defined.23 For the purpose of our analysis, BPD was categorized as present or absent.
At 54 months' CA, each child was tested with the Wechsler Preschool and Primary Scale of Intelligence-Revised (WPPSI-R, 1989).24 The WPPSI-R is an individually administered norm-referenced instrument that assesses the intellectual functioning of children (aged 3 years and 0 months through 7 years and 3 months) and provides three intelligence scores, a performance intelligence quotient (IQ), a verbal IQ, and a full scale IQ. All three IQ scores have a mean of 100 and a SD of 15. Ten subtests were administered; five performance subtests (object assembly, geometric design, block design, mazes, and picture completion) and five verbal subtests (information, comprehension, arithmetic, vocabulary, and similarities) were used for calculation of the total performance and verbal scores. IQ scores are derived from these scores. An attempt was made to administer the test to every child. If the child was delayed severely and unable to complete the test items, an IQ score of 40 was assigned for the WPPSI-R.
Although all children in the study have received neurodevelopmental follow-up, those children being raised in bilingual or non-English-speaking homes were excluded from the IQ data analysis, because the WPPSI-R was standardized with English-speaking persons. Thus, this test generates valid results within the population for which it was standardized. All children typically received the standard testing battery first, followed by other measures that may have been more clinically appropriate.
A routine neurologic examination was also part of the 54-month CA follow-up assessment.10 Assignment of cerebral palsy into classic groups was based on the presence of hypertonicity, hyperreflexia, and dystonic or spastic movement quality in the affected extremities. The groups were spastic diplegia, quadriplegia, and hemiplegia.
All demographic information was obtained from the primary care givers, and teams that performed the neurodevelopmental examinations remained blinded with regard to the subjects' previous medical histories. For the purpose of our analysis, the highest level of maternal education was categorized into one of four groups: less than a high school graduate, high school graduate, some college, and a college graduate or greater.
Categorical data with no expectation of a linear trend among groups (eg, gender) were analyzed using Fisher's exact test. Categorical data with an a priori expectation of a linear trend among groups were analyzed by the χ2 test for linear trend. The two-sample Wilcoxon rank-sum test was used for between-group comparisons of continuous-valued data. Logistic regression analysis was used to evaluate the effect on outcome (ie, IQ <70 vs IQ ≥70) of several possible predictive factors simultaneously. OR and two-sided 95% CIs were calculated based on the logistic model. An OR >1 implies that the odds of having a child with an IQ <70 increases with each progressive increase in the value of the predictive factor, whereas an OR <1 implies that the odds of having a child with an IQ <70 decreases with each progressive increase in the value of the predictive factor. For example, no IVH, grades 1 to 2 IVH, and grades 3 to 4 IVH were coded for purposes of analysis as −1, 0, and 1, respectively. Thus, an OR >1 would imply that the odds of having a child with an IQ <70 increases as one proceeds from no IVH to grade 1 to 2 IVH, and from grade 1 to 2 IVH to grade 3 to 4 IVH. All statistical analyses were performed using SAS software. All P values in this report are of the two-sided type.
Of the 505 infants admitted to our study protocols, 440 survived. Complete ultrasound data are available for 433. The mean birth weight of this population was 960 ± 172 g; the mean gestational age was 28.0 ± 2.0 weeks; 236 (55%) were boys.
Neurodevelopmental testing was available for 384 of the surviving children and is reported for the 257 subjects residing in English-speaking, monolingual households. Of the 257 (4%) subjects, 11 had moderate to severe VM, and 246 did not have VM. Representative cranial ultrasounds demonstrating VM and no VM at term are shown in Fig 1. Data for these two groups of children residing in English-speaking, monolingual households are found in Table 1 and are qualitatively similar to those for children residing in non-English-speaking, monolingual households. As shown, there were no differences in the gestational ages, number of male infants, Apgar scores at 1 and 5 minutes, and episodes of documented sepsis between the infants with and without VM. Children with VM were found to have had significantly lower birth weights than were children without VM (866 ± 152 g compared with 970 ± 179 g; P = .05). IVH was detected in all 11 (100%) children with VM, compared with 45 of the 246 (18%) children without VM (P < .001). In addition, 2 (18%) infants with VM had PVL, compared with 8 (3%) children without VM (P = .06). Similarly, BPD was diagnosed in 9 (82%) children with VM, compared with 120 (49%) infants without VM (P = .03). Analysis of risk for VM in this population demonstrated that both IVH and BPD were associated with VM. IVH was the stronger predictor of moderate to severe VM (OR: 36.6; 95% CI [8.0, 166.8]; P < .001), with the odds of having moderate to severe VM 36.6 times greater with each increase in IVH category (ie, from none to grades 1–2, and from grades 1–2 to grades 3–4). The OR with BPD as a predictor of VM was 4.73 (95% CI [1.0, 22.3];P = .05).
The neurodevelopmental data for these subjects demonstrated that 37 (14.4%) children had full scale IQ scores <70. Logistic regression models entering birth weight, gestational age, gender, Apgar score at 5 minutes, BPD, sepsis, PVL, VM, IVH category, and maternal education demonstrated that only maternal education (OR: 0.44; 95% CI [0.3, 0.7]; P < .001) and VM (OR: 19.0; 95% CI [4.5, 80.6]; P < .001) were important and independent predictors of full scale IQ <70 in our study population. An OR <1 for maternal education indicates that the odds of having a child with an IQ <70 decreases as the highest level of maternal education is increased. This model accurately predicted 82% of all children with IQ <70.
Neurodevelopmental scores for the 257 English-speaking, monolingual children are shown in Table 2. The 11 subjects with VM not only had a significantly lower full scale IQ than did the 246 children without VM (P = .02), but also had a significantly greater incidence of full scale IQ scores <70 (55% vs 13%; P = .002). Of the children with VM, 6 (55%) had full scale IQ scores <70, 1 (9%) had an IQ of 70 to 80, and 4 (36%) had IQ scores >80; in contrast, only 31 (13%) children without VM had IQ scores <70, 37 (15%) had scores of 70 to 80, and 178 (72%) received IQ scores >80 (P< .001). In addition, the performance IQ scores of the children with VM were more depressed (17.7 points lower), compared with the scores of children without VM (P = .009), than were the verbal IQ scores.
Although there was no difference between the ages of the mothers of the children with VM and those of the subjects without VM, the mothers of the children with VM had significantly more years of education (14.5 ± 2.5 years) compared with the education of mothers of the children without VM (13.1 ± 2.3 years; P = .04). Finally, the children with VM had cerebral palsy more frequently than did the subjects without VM (5/11 [45%] vs 17/237 [7%];P < .001). Of the 5 children with VM who had cerebral palsy, 2 (40%) were found to have spastic diplegia, 1 (20%) suffered spastic hemiplegia, and 2 (40%) had spastic quadriplegia. Similarly, of the 17 children without VM who had cerebral palsy, 8 (47%) had spastic diplegia, 3 (18%) were found to have spastic hemiplegia, and 6 (35%) suffered spastic quadriplegia.
The five WPPSI-R performance subtests, total performance score, the five verbal subtests, and the total verbal score are shown in Table 3. Children with VM performed more poorly than did children without VM on four of five performance subtests including object assembly, geometric design mazes, and picture completion. The most significant differences (P = .004) were in picture completion, a subtest in which the child must identify what is missing from pictures of common objects or events, and geometric design (P = .018), which includes vision recognition and vision perception motor integration tasks.
The two additional subtests with which the children with VM had difficulty (object assembly and mazes) both contain vision perception motor tasks. In contrast, the children with and without VM had similar scores on 4 of 5 verbal subscores, although there was a definite trend for children without VM to score higher, and this was reflected by the total verbal score. The children with VM scored lower (P = .049) on the arithmetic subtest, which requires the child to demonstrate an understanding of basic quantitative concepts. It begins with picture items, progresses to simple counting tasks, and ends with word problems.
Moderate to severe VM is a sign of both cortical and white matter injury in preterm infants.25–27 Recently, it has been associated with adverse outcome at school age. Whitaker reported that preterm subjects with VM and parenchymal involvement of hemorrhage performed significantly more poorly on measures of cognitive function at 6 years of age than did those children without these ultrasonographic findings. Similarly, the study by Stewart and Kirkbride8 of 14-year-old children with a history of preterm birth reported both a high incidence of VM detected by magnetic resonance imaging (41%) and poor school performance (22%). Stewart hypothesized that the white matter abnormalities found in her patients represent underlying alterations in hemispheric connectivity and thus provide a basis for the cognitive impairments in her study population. Unlike IVH and PVL, two well studied ultrasonographic findings in preterm infants, the pathophysiology of VM remains poorly defined.28 ,26
We used our cohort of very low birth weight preterm infants to test the hypothesis that moderate to severe cerebral VM at 40 weeks conceptional age is associated with adverse neurodevelopmental outcome and to examine the risk factors for this finding. These data demonstrated that 56% of infants with VM at term had full scale IQ scores <70 at 4.5 years of age, compared with 13% of children without VM (P = .002) despite an educational advantage for the mothers of the children with VM (P = .04). Children with VM at term suffered significant cognitive and motor handicaps at 4.5 years' CA, and our data show that the deficits were most pronounced in visual motor skills subtests. Of the children with VM, 45% suffered cerebral palsy, compared with 7% of the subjects without VM (P < .001). Finally, logistic regression demonstrated that VM is an independent and important predictor of IQ <70 in very low birth weight preterm infants.
The risk factors associated with this ultrasonographic finding included not only parenchymal involvement of IVH, but also low-grade hemorrhage and BPD. Although others have speculated that VM may be secondary to PVL, we noted an overall low incidence of PVL in our study patients.29 ,30 This finding may be secondary to our ultrasonographic techniques. Magnetic resonance imaging studies of our patients at 8 years of age are currently underway. We have reported previously that indomethacin protects against VM at term for very low birth weight infants with no evidence of IVH at 6 postnatal hours.31 Finally, although our study population of 11 (4%) children with VM is somewhat small to draw firm conclusions, we believe that the size of the comparison group of 246 subjects provides our study with a good deal of discriminatory power for detecting predictive factors for VM.
Oxygen deprivation is correlated with neurodevelopmental impairments in preterm infants.32 ,3 ,33 Infants with IVH experience prolonged severe depressions in cerebral blood flow.34 ,35BPD has also been considered by many scholars to represent a cause of oxygen deprivation to the developing brain.36–40 As many as 40% to 70% of all infants with birth weights <1000 g experience BPD,23 and 25% to 40% of infants with BPD suffer long-term neurodevelopmental handicap.28 Our data demonstrate that moderate to severe VM, a predictor of adverse outcome, is more common in children with BPD than in those without BPD.
Cerebral development in the human fetus is characterized by sequential periods of cellular proliferation, by migration of glia and neurons into appropriate cortical positions, and by the elaboration of synaptic connections with other cortical and subcortical regions of the brain.41–43 By 25 weeks of gestation, a time of birth for which the survival rate was 60% at Yale New Haven Hospital last year, almost all the developing cortical neurons have been generated44; the elaboration of axonal and dendritic arbors is at an active stage, and many synaptic contacts are being formed in the developing cortex.45 ,46
In the newborn rat model, the first 20 PNDs represent the period of rapid differentiation of axons and dendrites,47–49 and pharmacological insults such as ethanol and vitamin B6 deprivation have all been shown to decrease synaptic density and impair learning in this model.50–52 In addition, Laroia53 found significant decreases in cortical volumes in neonatal rats at PND 14 following a 3-hour hypoxic insult on PND 7. Newborn rats exposed to a model for chronic sublethal hypoxia had decreased cortical volumes and hemispheric white matter with significant VM.54 ,55O'Reilly reported that apical dendritic spine counts were markedly lower in rats raised in chronic hypoxia.56
VM secondary to the reduction in the volume of the white matter, as hypothesized by Stewart and Kirkbride8 suggests a profound effect on the pattern and level of cortico–cortical and cortico–fugal connectivity. Recent animal studies have begun to emphasize the critical role of the extrinsic circuitry interconnecting areas of the cerebral cortex in the highest levels of information processing.57 Clinical studies demonstrate that those infants with VM suffer not only abnormalities in the visual evoked response test,58 ,59 but also impaired visual motor performance.16 Our data provide additional evidence linking the findings of visual perceptual motor abnormalities with ventricular dilatation at term in preterm infants who are now 4.5 years' CA.
These data suggest that moderate to severe VM is a consequence of the vulnerability of the preterm brain. The presence of VM at term in very low birth weight infants is associated with adverse cognitive and motor outcomes at 4.5 years of age.
This work was supported by Grants NS 27116 and NS 35476 of the National Institute of Neurologic Disorders and Stroke and RR 06022 of the National Center of Research Resources.
We thank the following individuals from Yale University School of Medicine, Marjorene Ainley, BS, Susan DeLancy, MA, and Lisa Perry, MA; from Brown University, Terri Leach, MED, CAES, and Beth Dingley, BA; and from Maine Medical Center, Alison Milne, RNC, and June Gagnon, MA. We thank Giovanna Spinella, MD, for her scientific expertise.
- Received July 8, 1998.
- Accepted December 21, 1998.
Reprint requests to (L.R.M.) Department of Pediatrics, Yale University School of Medicine, 333 Cedar St, New Haven, CT 06511. E-mail:
- Guyer B,
- Strobino DM,
- Ventura SJ,
- MacDorman M,
- Martin JA
- Bhushan V,
- Paneth N,
- Kiely JL
- Robertson CMT,
- Hrynchyshyn GJ,
- Etches OC,
- Pin KS
- Piecuch R,
- Leonard C,
- Cooper B,
- Sehring S
- ↵Stewart A, Kirkbride V. Very preterm infants at fourteen years: relationship with neonatal ultrasound brain scans and neurodevelopmental status at one year. Acta Paediatr. 1996;416:44–47. Supplement
- Hack M,
- Friedman H,
- Fanaroff AA
- McCormick M,
- Workman-Daniels K,
- Brooks-Gunn J
- Klebanov P,
- Brooks-Gunn J,
- McCormick M
- Krishnamoorthy KS,
- Shannon DC,
- DeLong GE,
- Todres ID,
- Davis KR
- Whitaker AG,
- Feldman JF,
- Rossem RV,
- et al.
- Ment LR,
- Oh W,
- Ehrenkranz RA,
- Philip AGS
- ↵Wechsler D. Manual for the Wechsler Intelligence Scale for Children. 3rd ed. San Antonio, TX: Psychological Corp; 1991
- Volpe JJ
- Perlman JM,
- Risser R,
- Broyles RS
- ↵Ment LR, Vohr B, Oh W. Neurodevelopmental outcome at 36 months CA of preterm infants in the Multicenter Indomethacin IVH Prevention Trial. Pediatrics. 1996;98
- ↵Volpe JJ. Brain injury in the premature infant. Pediatr Res. 1990;27:28–33. Supplement
- Volpe JJ,
- Herscovitch P,
- Perlman JM
- Poets CF,
- Stebbens VA,
- Richard D,
- Southall DP
- Bourgeois JP,
- Jastreboff PJ,
- Rakic P
- O'Reilly J,
- Schartz M,
- Haddad G
- ↵Goldman-Rakic PS. Prefrontal cortex revisited: a multiple memory domain model of human cognition. In: Caminiti R, Hoffman K-P, Lacquaniti F, Altman J, eds. Vision and Movement Mechanisms in the Cerebral Cortex. Strasbourg, Germany: Human Frontier Science Program; 1996:162–172
- Ehle A,
- Sklar F
- McLone D,
- Czyzewski D,
- Raimondi A,
- Sommers R
- Copyright © 1999 American Academy of Pediatrics