School-Age Outcomes of Very Low Birth Weight Infants in the Indomethacin Intraventricular Hemorrhage Prevention Trial
Objective. The cohort consisted of 328 very low birth weight infants (600–1250 g birth weight) who were enrolled in the low-dose prophylactic indomethacin prevention trial and were intraventricular hemorrhage (IVH) negative at 6 postnatal hours. The objective was to determine the effects of both IVH and indomethacin on cognitive, language, and achievement performance at 8 years of age.
Methods. The cohort was divided into 4 subgroups for analysis: indomethacin plus IVH, indomethacin no IVH, saline plus IVH, and saline with no IVH. The children were evaluated prospectively at 8 years of age with a neurologic assessment, history of school performance, and a battery of cognitive, academic, behavioral, and functional assessments.
Results. Children in both IVH groups had more cerebral palsy; more hearing impairment; lower daily living skills scores; lower IQ, vocabulary, and reading and mathematics achievement test scores; and greater educational resource needs. With logistic regression analyses grade 3 to 4 IVH, periventricular leukomalacia and/or ventriculomegaly, male gender, maternal education, and language spoken in the home contributed to outcomes. No effects of indomethacin or gestational age were identified.
Conclusions. Although biological factors including IVH, ventriculomegaly, and periventricular leukomalacia contribute significantly to school age outcomes among very low birth weight survivors at 8 years of age, social and environmental factors including maternal level of education and primary language spoken in the home are also important contributors to outcome.
Preterm birth is known to be associated with an increased rate of neurodevelopmental handicap.1–4 One of the major contributions to adverse neurologic and cognitive outcome is intraventricular hemorrhage (IVH), particularly grade 3 to 4 IVH.5,6 Indomethacin, an inhibitor of prostaglandin synthesis, has been shown to decrease the incidence of IVH. In our previous reports of this study cohort, we demonstrated a decreased incidence of IVH,7–9 particularly severe grade 3 to 4 IVH, and a trend for improved survival in infants who were exposed to low-dose prophylactic indomethacin.8 Evaluation of this cohort at 4.5 years of age revealed similar rates of cerebral palsy (CP) in the 2 study groups.10 However, a modest cognitive benefit was identified for the children in the indomethacin group, and at 6 years we reported that indomethacin may preferentially protect language function. Because very low birth weight (VLBW) infants are known to have increased academic difficulties at school age, our objective was 1) to determine the effects of both IVH and indomethacin on cognitive, language, and school age performance at 8 years of age and 2) to assess the predictors of cognitive, language, and school age impairments at 8 years of age. It was hypothesized that the study children (all of whom were IVH negative around 6 hours of age) who did not develop IVH would demonstrate better cognitive and language skills and fewer learning difficulties at 8 years of age than children who developed IVH. A second hypothesis was that there would be beneficial effects of indomethacin on language performance.
The clinical studies were conducted at Women and Infants’ Hospital (Providence, RI), Maine Medical Center (Portland, ME), and Yale New Haven Hospital (New Haven, CT). The protocols and procedures described below were reviewed and approved by the institutional review boards of the 3 participating institutions.
During the time interval September 1, 1989, through August 30, 1992, parental informed consent was obtained for 505 infants with a birth weight of 600 to 1250 g admitted by 6 hours of age to participate in a randomized, prospective trial designed to determine whether the early administration of low-dose indomethacin would prevent IVH in this patient population.8,9 To evaluate the brain for hemorrhage, all infants enrolled in the study were first examined using cranial echoencephalography (ECHO) between 5 hours and 11 hours. Of the 505 enrolled infants, 431 did not have IVH around 6 hours and were randomized to the primary IVH prevention trial. These subjects are the basis of this study. Of the original cohort, 328 of the 384 long-term survivors (85%) were evaluated at 8 years of age. There were 165 children in the indomethacin group and 163 in the saline group. The 56 children who were not evaluated did not differ from the 328 evaluated at 8 years of age in birth weight, gestation, IVH status, bronchopulmonary dysplasia (BPD), or level of maternal education. The subjects were divided further into subgroups based on whether they developed IVH or did not.
Subsequent scans to evaluate the brain for hemorrhage and/or ischemic changes were performed at 24 and 48 hours after the first cranial ECHO, and on postnatal days 4, 5, 7, 14, and 21, and at 40 weeks postmenstrual 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 reported8: grade 1, blood in the periventricular germinal matrix regions; grade 2, blood within the lateral ventricular system without ventricular dilation; grade 3, blood within and distending the lateral ventricles; and grade 4, blood within the ventricular system and parenchymal involvement. IVH was categorized into 3 groups: none, grades 1 to 2, and grades 3 to 4. Ventriculomegaly (VM) was assessed on the ECHO studies performed at 40 weeks’ conceptional age (or, if not available, 21 days of age). For the purpose of our report of VM, 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) were included.11 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.11
As previously described, all infants underwent gestational age assessment using a modification of the Ballard scale.12 Prenatal, 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 BPD if he/she both required oxygen supplementation and had an abnormal chest radiograph at 28 days of life, as previously defined.13 For the purpose of our analysis, BPD was categorized as present or absent.
Assessments were performed at 8 years of age by testers who were blinded with regards to the child’s study medication status. The following tests were administered: the Wechsler Intelligence Scale for Children-Third Edition (WISC-III)14 is an individually administered norm-referenced instrument for assessing the intellectual functioning of children ages 6 years 10 months through 16 years 11 months. The test provides a verbal, performance, and full-scale IQ.
The Peabody Picture Vocabulary Test-Revised (PPVT-R)15 is a multiple-choice test that measures receptive vocabulary development and listening vocabulary in a format that requires no verbal response from the child. The child is required to identify a stimulus word from a selection of 4 pictures.
The “Tapping Test”16 is an inhibitory task in which the child and experimenter take turns holding a wooden dowel. In the first condition, the child is asked to imitate what the examiner does, tapping once when the examiner taps once, and tapping twice when the examiner taps twice. In the second condition (contrasts), when the experimenter taps twice, the child taps once, and when the experimenter taps once, the child taps twice. Two practice trials are given, and the child is corrected for incorrect responses. The training procedure is repeated until the child is correct on 2 consecutive trials (single and double tap). Then, 16 trials are presented with the order of 1 and 2 taps randomized.
The Vineland Adaptive Behavior Scales Survey Form17 is a semistructured interview administered to parents to assess the child’s adaptive functioning in the areas of communication, daily living skills, and socialization.
The Child Behavior Checklist (CBCL)18 is a widely used standardized instrument designed to assess the social competencies and behavior problems of children 4 to 18 years of age. Parents complete questions regarding their child’s performance in sports, classroom activities, chores, and the quality of relationships with friends and family. The Vineland Adaptive Behavior Scales and the CBCL are administered to a parent by a blinded examiner.
The Peabody Individual Achievement Test-Revised (PIAT-R)19 is a battery of academic achievement tests for children ages 5 years 0 months to 18 years 11 months. The reading recognition, mathematics, and spelling subtests were administered.
Demographic questionnaires, which included information on household composition, languages spoken in the home, educational level of the biological or adoptive/foster parent, school placement of the child, resource utilization, and need for medications, were administered to the parent/caregiver.
A standard neurologic examination was performed by the pediatric neurologist/developmental pediatrician at each site. The neurologic examination included determinations of height, weight, and occipitofrontal head circumference, assessment of visual field testing, pupillary function, eye movements, and facial strength. Tone, strength, reflexes, and cerebellar function were determined. A determination of normal, suspect, or abnormal neurologic status was made. Abnormal or suspect assessments were secondarily assessed for the presence of microcephaly, spastic diplegia, hemiplegia, or quadriplegia.
We completed 2-group analyses (indomethacin vs saline) to identify any main effects of indomethacin on outcome. This was followed by 2-way analysis of variance to identify the effects of both indomethacin and IVH on outcome. 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. Analyses of variance were used for between-group comparisons of continuous-valued data. Logistic regression analyses with backward elimination were run to predict IQ scores at 8 years of age. All statistical analyses were performed using SAS software (SAS Institute, Inc, Cary, NC). All P values in this report are of the 2-sided type. Multiple regression analyses were run to predict the standard scores for the WISC-III, performance IQ, verbal IQ and full-scale IQ, PPVT-R, and the PIAT-R tests. Independent variables were indomethacin versus saline, male versus female, gestational age, grade 3 to 4 IVH versus no grade 3 to 4 IVH, VM + PVL versus none, less than high school graduate versus more than or equal to high school graduate, and English monolingual versus bilingual household. Indomethacin was retained in the models by force.
Twenty (12%) of the 165 infants in the indomethacin group and 28 (17%) of the 163 infants in the saline group (P = .07) evaluated at 8 years of age had IVH. Characteristics of the 4 subgroups (indomethacin plus IVH, indomethacin with no IVH, saline plus IVH, saline with no IVH) evaluated at 8 years of age are shown in Table 1. The indomethacin infants had a significantly lower gestational age than the placebo group, P = .008. In addition, children with IVH had higher rates of BPD, P = .002. All other characteristics were similar among groups.
Neurologic and neurosensory status of the 4 groups of children at 8 years of age are shown in Table 2. Ninety-two percent of the indomethacin group and 94% of the saline group who did not develop IVH were neurologically normal. Children in both the indomethacin and saline groups with IVH were more likely to have CP, 32% versus 6% and 20% versus 5% than the children who did not develop IVH (P < .0001). A need for glasses did not differ among groups. Hearing impairment requiring amplification was also more prevalent in children in both the indomethacin and saline groups with IVH compared with the children without IVH, respectively (P = .02).
None of the 2-group analyses for the cognitive tests were significant. Cognitive test scores for the 4 study groups at 8 years are shown in Table 3. Children in both the indomethacin and saline groups who did not develop IVH had significantly higher WISC-III, verbal IQ, performance IQ, full-scale IQ, and PIAT-R standard scores for reading recognition, reading comprehension, and math. They also had fewer contrasts and total number of errors on the Tapping Tests. Analysis of group differences on the PPVT-R revealed significant positive indomethacin effects and negative IVH effects. Only 8% of children in the indomethacin with no IVH group had vocabulary skills <70 compared with 18% of children in the saline with no IVH group and 25% of children in the 2 IVH-positive groups (P = .036).
Maternal perception of their child’s functional status on the Vineland and their behavior on the CBCL is shown in Table 4. Although there were trends for lower scores in all domains on the Vineland for children with IVH, the findings were significant only for the daily living skills domain. Children with IVH in both the indomethacin and saline groups had lower scores than children without IVH. In addition, children in the indomethacin groups had more scores in the daily living skills domain lower than 70 than children in the saline group. There were no significant group differences identified on the CBCL.
Table 5 shows that resource utilization at 8 years of age was significantly higher for children with IVH than those without. Seventy-nine percent of children in the indomethacin plus IVH group and 62% in the saline plus IVH group were receiving some academic resource support or therapy compared with 49% and 47% of the children without IVH (P = .009). Children with IVH were receiving twice as many individual therapies (P < .0001), were more likely to be in a self-contained classroom (P < .01), and were more likely to be receiving speech or language therapies (P < .05). Repeating a grade was similar among groups. Rates of facilitated communication, behavior counseling, and rehospitalization between the 6-year and 8-year visit and use of long-term medications did not differ among groups.
Table 6 shows the results of the 7 regression models to predict the standard scores for the performance IQ, verbal IQ, full-scale IQ, PPVT-R and PIAT-R reading recognition, reading comprehension, and mathematics achievement tests. Indomethacin treatment group and gestational age at birth did not contribute independently to any of the models. IVH 3 to 4 was negatively associated with all IQ scores and the PPVT-R scores providing 5.9%, 5.8%, 7.0%, and 5.2% of the variance, respectively. IVH 3 to 4 did not provide an independent contribution to the models for the PIAT-R tests. VM and/or PVL were negatively associated with all outcomes contributing >3% variance to performance IQ, full-scale IQ, and reading comprehension (3.5%, 3.0%, and 3.7%, respectively). Male gender was negatively associated with both the reading recognition and reading comprehension scores. Of the 2 environmental factors, low maternal education was associated negatively with all outcomes contributing the greatest amount of variance to verbal IQ, PPVT-R scores, and reading recognition scores (8.8%, 7.7%, and 8.9% of variance, respectively). Primary language other than English in the home was negatively associated only with verbal IQ and PPVT-R scores and contributed small amounts of variance (1.1% and 1.5%) to the models. Total variance within the models to predict the 8-year outcomes ranged from 9% to 18%.
Our objective was to investigate the relative contribution of both IVH and prophylactic low-dose indomethacin to the cognitive, language, and academic outcomes of VLBW infants at 8 years of age. Therefore, in this report we completed 4-group rather than 2-group analyses. It was not unexpected that the study children who developed IVH were more likely to have a diagnosis of BPD and a slightly lower gestational age, indicating they were sicker during their neonatal hospitalization.20 IVH, which developed after 12 hours of postnatal age in both the indomethacin and saline subgroups, had negative effects on all study outcomes. This is clearly demonstrated in the rates of cerebral palsy (CP), which were 32% and 20% for indomethacin and saline groups with IVH compared with only 6% and 5% for the respective groups without IVH. This finding represents a 4- to 5-fold increase among children with IVH and supports the need to continue to investigate interventions to prevent IVH in VLBW infants.
We had previously shown no adverse effects of indomethacin on outcome in this cohort at 3 and 4.5 years of age.10,21 This is consistent with our findings at 8 years of age. Bivariate analyses indicated that children in both the indomethacin and saline groups who did not develop IVH were doing well and had mean verbal IQ scores in the average range 95.7 + 18 and 95.2 + 17 with only 7% of scores >2 standard deviations below the mean. Trends were similar for performance and full-scale IQ scores. Rates of poor performance (scores <70) were significantly higher in the children with IVH for IQ, vocabulary, achievement tests, and the tapping tests. Twenty-five percent of children with IVH in both the indomethacin and saline groups and 18% of the children in the saline with no IVH group had scores <70 on the PPVT-R. In contrast, only 8% in the indomethacin with no IVH group had PPVT-R scores >2 standard deviations below the mean, suggesting a beneficial effect of indomethacin on vocabulary skills. This was consistent with our previous findings of a possible beneficial effect of indomethacin on language.
Although we observed some trends on the Vineland, we did identify significant negative effects of both IVH and indomethacin on maternal perception of daily living skills. The IVH effects are not unexpected because of the higher rates of CP in the IVH group. McCormick et al22 had reported that half of extremely low birth weight children had a health condition or limitation in an activity of daily living. In our cohort rates of function >2 standard deviations below the mean ranged from 14% for the saline groups to 50% for the indomethacin plus IVH group. The 50% rate in the indomethacin plus IVH group is partially explained by their 32% rate of CP. We also see that there was a trend for mothers of children in the indomethacin with no IVH-positive group to identify more thought problems (24%) and attention problems (24%) in their children on the CBCL. Others studies have shown an increase in behavior disorders among children with central nervous system lesions.23,24
A number of studies25–31 have reported that close to 50% of VLBW children require some educational resources at school age. This is consistent with our study results for the children without IVH (49% and 47%). However, the rates were considerably higher for the indomethacin plus IVH (79%) and saline plus IVH (62%) groups. The children with IVH also had significantly greater needs for numbers of individual resources needed, a self-contained classroom, and speech/language therapy. Within the 2 IVH-positive groups, 31% and 35% of children were receiving 4 or more separate therapies compared with 9% and 11% of children without IVH.
Our intention with the multivariate analyses was to enter robust biological and social environmental predictors of neurodevelopmental outcome. We also wished to identify any positive or negative effects of indomethacin. Therefore, indomethacin was retained in the models for all steps by force. Logistic regression analyses revealed that both the biological factors (IVH 3–4 and VM and/or PVL) and environmental factors (maternal education and language spoken in the home) were important predictors of our outcomes. IVH 3 to 4 had significant negative effects on verbal, performance, and full-scale IQ. This is consistent with prior reports.8–11,21,32 However, IVH 3 to 4 did not contribute to the models to predict performance on the achievement tests for reading and mathematics. The biological predictor of poor performance on achievement tests was the presence of VM and/or PVL. This would suggest that the more discrete areas of learning and achievement for reading and mathematics are impacted to a greater extent by periventricular white matter lesions, VM, and cortical atrophy.32 We previously reported the negative effects of both VM and PVL on cognitive test scores.11,32 Our previous report on a subgroup of this cohort demonstrated an association between low cortical volumes and low cognitive scores.32 Although all the models indicate that VM and/or PVL have a negative impact on IQ, vocabulary, and achievement tests, the amount of variance contributed to the model was low (1.6% to 3.7%). IVH contributed negatively to the IQ and vocabulary scores contributing between 5.1% and 5.9% variance. Although others have shown that lower birth weight or gestational age are associated with decreased survival and increased sequelae,33–35 our multivariate analysis at 8 years of age indicated that gestational age did not contribute to the models, suggesting that with increasing age environmental factors may be stronger predictors than some of the biological factors.
Numerous social and environmental factors are known to impact on the outcomes of infants,36–39 and we identified maternal education and residing in a bilingual household as our 2 environmental predictors. We also added gender to the model because male children are at increased risk of developmental delay,37,38 and male gender was negatively associated with both the reading recognition and reading comprehension achievement tests. However, maternal education less than high school impacted negatively on all the study outcomes. In fact, maternal education less than high school predicted most of the variance in the models and was an important modifier contributing to all outcomes. Low maternal education contributed significant negative variance especially to verbal IQ (8.8%), the PPVT-R (7.6%), PIAT-R Reading Comprehension (8.9%), and PIAT-R Mathematics (6.7%). This is consistent with other reports demonstrating that social risks are good predictors of verbal skills.39 Residing in a bilingual or non-English monolingual household contributed small amounts of negative variance only to the verbal IQ and PPVT-R at 8 years of age. Residing in a bilingual or non-English monolingual household had effects similar to, but weaker than, low maternal education. Although bivariate analysis of indomethacin suggested a possible beneficial effect on vocabulary skills, this was not confirmed with multivariate analyses. We recognize that many other biological and environmental factors contribute to outcome, but evaluation of multiple factors would require an alternate study protocol and a larger sample size. Our analyses have focused on evaluating specific predictors in a longitudinal well-defined cohort.
Grade 3 to 4 IVH remains an important predictor of poor outcome among children with a birth weight of 600 to 1250 g at school age. In addition, environmental factors (maternal education and bilingual household) play an equally important role in predicting the neurodevelopmental and educational outcomes of VLBW children at school age. The IVH-associated increased needs for child health and educational services represent an important public health concern. Efforts to identify and facilitate management interventions to prevent IVH, PVL, and VM, and family-centered interventions to enhance development are needed to improve the school age outcomes of VLBW infants.
This work was supported by NS27116 and the National Center of Research Resources RR 06022.
- ↵Doyle LW, Casalaz D. Outcome at 14 years of extremely low birthweight infants: a regional study. Arch Dis Child Fetal Neonatal Ed.2001;85 :F159– F164
- ↵Perlman MM, Rollins N, Burns D, Risser R. Relationship between periventricular intraparenchymal echodensities and germinal matrix hemorrhage in the very low birth weight neonate. Pediatrics.1993;91 :474– 480
- ↵Volpe JJ. Brain injury caused by intraventricular hemorrhage: is indomethacin the silver bullet for prevention? Pediatrics.1994;93 :673– 677
- ↵Ment LR, Oh W, Ehrenkranz RA, et al. Low dose indomethacin and prevention of intraventricular hemorrhage: a multicenter randomized trial. Pediatrics.1994;93 :543– 550
- ↵Ment LR, Vohr B, Allan W, et al. Outcome of children in the indomethacin intraventricular hemorrhage prevention trial. Pediatrics.2000;105 :485– 491
- ↵Allan WC, Philip AGS. Neonatal cerebral pathology diagnosed by ultrasound. Clin Perinatal.1985;12 :195– 216
- ↵Wechsler D. Manual for the Wechsler Intelligence Scale for Children. 3rd ed. San Antonio, TX: Psychological Corp; 1991
- ↵Dunn LM, Dunn LM. Peabody Picture Vocabulary Test- Revised. Circle Pines, NM: American Guidance Service; 1981
- ↵Lezak M. Neuropsychological Assessment. 3rd ed. New York, NY: Oxford University Press; 1995:672
- ↵Sparrow SS, Balla DA, Cicchetti DV. Vineland Adaptive Behavior Scales. Circle Pines, MN: American Guidance Service; 1984
- ↵Achenbach TM. Manual for the Child Behavior Checklist/4-18 and 1991 Profile. Burlington, VT: University of Vermont Department of Psychiatry; 1991
- ↵Markwardt FC. Peabody Individual Achievement Test-Revised Manual. Circle Pines, MN: American Guidance Service; 1989:1–243
- ↵Ment LR, Vohr B, Oh W, et al. Neurodevelopmental outcome at 36 months corrected age of preterm infants in the Multicenter Indomethacin Intraventricular Hemorrhage Prevention Trial. Pediatrics.1996;98 :714– 718
- ↵McCormick M, Brooks-Gunn J, Workman-Daniels K, Turner J, Packmah G. The health and developmental status of very low-birth weight children at school age. JAMA.1992;49 :783– 790
- Msall ME, Buck GM, Rogers BT, Catanzaro NL. Kindergarten readiness after extreme prematurity. Am J Dis Child.1992;146 :371– 375
- Byrd RS, Weitzman ML. Predictors of early grade retention among children in the United States. Pediatrics.1994;93 :481– 487
- Hack M, Klein N, Taylor HG. School-age outcomes of children of extremely low birth weight and gestational age. Semin Neonatol.1996;1 :227– 288
- ↵Peterson BS, Vohr B, Staib LH, et al. Regional brain volume abnormalities and long-term cognitive outcome in preterm infants. JAMA.2000;18 :284:1939– 1947
- ↵Weisglas-Kuperus N, Baerts W, Smrkovsky J, Sauer PJJ. Effects of biological and social factors on the cognitive development of very low birth weight children. Pediatrics.1993;92 :658– 665
- ↵Vohr BR, Wright LL, Dusick AD, et al. Neurodevelopmental and functional outcome of extremely low birth weight (ELBW) infants in the NICHD Neonatal Network 1993–1994. Pediatrics.2000;105 :1216– 1226
- Copyright © 2003 by the American Academy of Pediatrics