Objective. The aim of this study was to ascertain the early developmental status of children who have a history of newborn encephalopathy.
Methods. A longitudinal follow-up was conducted of a population-based, case-control study of children born in Western Australia between June 1993 and December 1996. The study included 276 term children (≥37 weeks’ gestation) with moderate or severe newborn encephalopathy and 564 unmatched term control subjects. The Griffiths Mental Development Scales was used to ascertain developmental status and a General Quotient (GQ) score. Outcome measures were the Griffiths developmental subscales, GQ, diagnosis of cerebral palsy, and mortality.
Results. Thirty-four patients and 1 control subject died before reaching assessment. Between June 1994 and December 1999, 195 (81%) eligible patients and 445 (79%) eligible control subjects were assessed. Statistically significant differences were found between patients and control subjects for GQ and all developmental subscales. Overall, 39% of patients had a poor outcome as defined by death, cerebral palsy, or a significant degree of developmental delay, compared with 2.7% of control subjects. Furthermore, 62% of those with severe encephalopathy had a poor outcome compared with 25% of those with moderate encephalopathy. Patients with a history of seizures were 3 times more likely to develop cerebral palsy than patients without. Overall, 28 (10.1%) of patients have cerebral palsy.
Conclusions. These data provide important prognostic information regarding survival and serious disability and indicate that newborn encephalopathy places children at significant risk of developmental delay by their second year. These findings also suggest that comprehensive clinical and educational assessments are required to enable appropriate educational provisions as these infants approach school entry.
Few long-term studies of the outcomes after newborn encephalopathy have been conducted. Of those conducted, the majority were not population-based, most concentrated on encephalopathy associated with “birth asphyxia,” and others included only infants with neonatal seizures. Few studies have been concerned with outcomes other than cerebral palsy and death.1–12 With notable exceptions, other disabilities such as cognitive impairment and developmental delay have not been considered or have been reported only for infants with hypoxic ischemic encephalopathy.1,8,11–13 We undertook a case-control study of moderate and severe newborn encephalopathy with recruitment from 1993 to 1996.14,15 This was the first population-based study of newborn encephalopathy using a broad clinical definition. We subsequently followed the patients and control subjects longitudinally. The aim of the first stage of the follow-up, reported here, was to ascertain the developmental status of the children during their second year of life; mortality and cerebral palsy are also reported.
The methods of the original case-control study have been reported in detail elsewhere,14,15 although we subsequently extended the study by continuing to recruit additional patients and control subjects. Briefly, all 276 cases of moderate or severe newborn encephalopathy in term infants, born in Western Australia between June 1993 and December 1996, who fulfilled the criteria in Fig 1, were prospectively enrolled irrespective of presumed cause. The newborn encephalopathy was graded as moderate or severe according to criteria modified from Sarnat and Sarnat.16 Patients who were defined as having severe newborn encephalopathy were those who fulfilled 1 or more of the following criteria: ventilation for >24 hours; 2 or more anticonvulsant treatments required; comatose or stuporous; died in the neonatal period. The remaining case infants were defined as having moderate encephalopathy.
The control subjects were 564 term live births (97% response) who were born during the study period in the Perth Metropolitan area and who were randomly selected using the method described by Draper et al.17 The control subjects are a robust random sample of term live births in the study population.15
Patients and control subjects were assessed using the Griffiths Mental Developmental Scales (birth-8 years).18 The majority were assessed between ages 1 and 2 years. The assessments were conducted by 2 trained assessors who were blind to the case-control status of the child. A small number of assessment results were taken from routine Griffiths assessments performed for clinical purposes at the psychology department at the tertiary children’s hospital. Of those eligible, 29% of patients and 31% of control subjects were assessed at home.
The Griffiths General Quotient (GQ) score data for the control subjects were normally distributed. The population mean and standard deviation (SD) for the GQ score were calculated for the control subjects. A cutoff of 2 SD below the control population mean GQ score was used to identify patients and control subjects with a clinically significant degree of developmental delay. The GQ score and Griffith subscale data for the patients were markedly skewed to the left (Fig 2). Median values were therefore compared between patients and control subjects using the Wilcoxon rank sum test for nonparametric data. Odds ratios and their 95% confidence intervals were estimated using the method of Clayton and Hills.19 The formal level of statistical significance was taken as P < .05.
Deaths were identified from data collected on a weekly basis from the Western Australian Registrar General. Cases of cerebral palsy were identified from the Western Australian Cerebral Palsy Register.20,21
Overall, 276 cases with newborn encephalopathy and 564 control subjects were enrolled in the original case-control study. Of these, 34 patients and 1 control subject died before reaching the Griffiths assessment age. Thus, 242 patients and 563 control subjects were eligible for assessment. By December 31, 1999, 190 cases had been assessed using the Griffiths Mental Development Scales, 4 had been assessed using instruments more suited to the child’s clinical condition, eg, blindness, and 1 case had such a severe disability that he could not be assessed. This gave a follow-up fraction of 81% for the eligible patients. The alternative assessments for patients included the Reynell Zinkin Developmental Scales for Young Visually Handicapped,22 the Bayley Scales of Infant Development II,23 the Stanford-Binet Intelligence Scale (4th edition),24 and a Rural Pediatric Service Assessment. Results indicated that 2 of the patients were found to be developing normally, and 2 were detected as having developmental delay. Of the 563 eligible control subjects, 443 had been assessed with the Griffiths Scales, and 2 had been assessed with other instruments, giving a follow-up fraction of 79% (Table 1). The Sewell Early Education Developmental Profile of Development25 and a Health Service Assessment were the alternative assessments used for the 2 control subjects, both of whom were found to be developing normally. A Griffiths assessment was not conducted on 47 eligible patients (19%) and 118 eligible control subjects (21%). Of note is that 7 patients (3%) and 20 control subjects (4%) are still awaiting assessment because they now live either overseas or in rural Western Australia at distances of up to 3300 km from Perth.
The median chronological age at assessment was 16.4 months for the patients and 16.0 months for the control subjects. Sixty percent of the patients and 53% of the control subjects assessed were boys. The birth characteristics of patients and control subjects have been published elsewhere.14,15
Griffiths Assessment Results for All Eligible Infants
The results of the Griffiths assessments are given in Table 2 and Fig 2. Statistically significant median differences between the patients and control subjects in the GQ score and the 5 subscales ranged from 6.0 to 8.0 points. These differences represent a median developmental age deficit of 1.5 to 2.5 months for the patients compared with control subjects; the largest deficit was in the speech and hearing subscale. The population mean GQ score estimated from the control subjects was 113 (SD: 9.3). Of the 190 patients assessed, 23.3% had GQ scores below the cutoff of 2 SD below the population mean (94.5 points), compared with 2.5% of the control subjects (P < .001).
Griffiths Assessment Results Excluding Infants With Cerebral Palsy
Overall, 28 (10.1%) of the 276 patients with newborn encephalopathy and none of the control subjects were known to have developed cerebral palsy when this analysis was performed. Of the 190 patients assessed, 21 (11%) are known to have cerebral palsy and 2 have subsequently died. Of the 7 remaining patients with cerebral palsy who were not assessed, 3 died before reaching assessment age, 1 refused an assessment, 2 withdrew from the study completely, and 1 was blind and therefore had an alternative assessment. To date, none of the control subjects have developed cerebral palsy.
Table 3 and Fig 2 give the results for the patients and control subjects who were assessed having excluded those patients who developed cerebral palsy. Compared with the results for all of the cases, these results were attenuated; however, statistically significant median differences between the patients and control subjects were seen. These ranged from 3.7 to 6.2 points, which represent an overall median developmental age deficit of 1.0 to 1.5 months. The largest deficit was again seen in the speech and hearing subscale. Of the 169 cases assessed, 15.5% had a GQ score below the significant developmental delay cutoff, compared with 2.5% of the control subjects (P < .001).
Griffiths Assessment Results Excluding Infants With Cerebral Palsy and Contributory Medical Conditions
Seventeen patients and 4 control subjects had medical conditions that were thought to be likely to contribute to a developmental deficit. These conditions included organ failure, metabolic disorders, hydrocephalus (requiring shunt), profound deafness, Vater syndrome, muscular dystrophy, hypothyroidism, and global developmental delay. These infants and those with cerebral palsy were then excluded from the analysis. Table 4 and Fig 2 illustrate the Griffiths results after their exclusion. Compared with when all of the children were assessed, again, the median differences were attenuated. However, statistically significant differences between the patients and control subjects ranged from 2.9 to 5.7 points, which represent an overall median developmental age deficit of 0.5 to 1.5 months. Again, the largest deficit was seen in the speech and hearing subscale. Of the remaining 152 patients assessed, 9.3% had GQ scores that indicated a significant developmental delay, compared with 2.5% of the remaining 439 control subjects (P < .001).
Figure 3 summarizes the early childhood outcomes for the patients and control subjects. Thirty-nine percent of the patients had a poor outcome in early childhood as defined by death, cerebral palsy, or significant developmental delay, compared with 2.7% of the control subjects. The 47 patients and 118 control subjects who were eligible for assessment and were not assessed are known to be alive and do not have cerebral palsy; however, their developmental status is not known. If we assume that all of the unassessed patients and control subjects had a GQ score above the cutoff, then, overall, one third of the patients would have a poor outcome compared with only 2.1% of the control subjects (P < .001).
Table 5 compare the outcomes for all cases by the grade of their newborn encephalopathy. Overall, 25% of the patients with a history of moderate encephalopathy had a poor outcome in early childhood compared with 62% of patients with a history of severe encephalopathy. Death in the neonatal period was one of the defining features of severe encephalopathy, and 24 patients died in this period. Having excluded these patients, 47% of the remaining patients with severe encephalopathy had a poor outcome. Compared with the patients with moderate encephalopathy, the patients with severe encephalopathy were more than twice as likely to have a GQ score below the cutoff, indicating significant developmental delay. This result was statistically significant when all of the assessed cases were included. Once the children with cerebral palsy and contributory medical conditions were excluded, the odds ratio estimates were similar, however, as only a small number of the severe cases remained, the results were no longer statistically significant.
The outcomes for the patients were also compared by whether they had experienced seizures as part of their encephalopathy (Fig 5). Patients with a history of seizures were more than 3 times more likely to develop cerebral palsy (16%) compared with those without a history of seizures (6%; odds ratio: 3.4 [95% confidence interval: 1.14, 10.2]; P = .04). Other outcomes, however, were remarkably similar as illustrated by the proportions of patients with and without a history of seizures whose GQ scores fell below the significant developmental delay cutoff value (Table 5).
The findings from our population-based study indicate that newborn encephalopathy places infants at significant risk of developmental delay by the second year of life. We found differences in all areas of development, as assessed by the Griffiths Mental Development Scales, which were both statistically and clinically significant. Of note, the largest deficits were seen in speech and hearing, which are crucial areas for all aspects of development and learning.
We excluded the cases with cerebral palsy or conditions known to be associated with significant developmental delay to answer the question most frequently asked by parents: “If my child does not have cerebral palsy, then what is the outlook?” Once these children had been excluded, the differences in development and GQ score between the patients and control subjects were still evident, although not as extreme. Nevertheless, 9% of the remaining patients still had a GQ score that indicated that they had a clinically significant degree of developmental delay.
We were not able to assess all of the patients and control subjects who were eligible for assessment largely because of the vast distances involved in Western Australia. However, even assuming the best-case scenario for those whom we did not assess, the infants with a history of newborn encephalopathy were at a significant developmental disadvantage. Of greatest concern is the impact that the delay will have on future learning and educational attainment. Furthermore, given that we assessed the children at a relatively young age, it is likely that more subtle disabilities and difficulties will declare themselves later, especially at school.
Most previous studies of the long-term outcomes after newborn encephalopathy were conducted some time ago and were concerned mainly with mortality and gross motor impairment.2,3,6 A notable exception is the study by Robertson et al,1 which described other neurodevelopmental outcomes. Nevertheless, the study was restricted to neonates with perinatal asphyxia.1 Other studies in the literature have been smaller and have mainly described infants who have experienced seizures alone8–10,12 or intrapartum hypoxia.5–7,11 Most recently, the 1-year outcomes from a prospective cohort study of newborn encephalopathy in Kathmandu, Nepal, have been published.13 In contrast to our study, the authors included children with mild encephalopathy and excluded those with neonatal sepsis, congenital malformations, or primary hypoglycemia. By 1 year of age, 44% of the 131 cases had died; 18% had severe impairments, mostly quadriplegic cerebral palsy; and 2% had minor impairments. They also described an excess mortality (17%) for mild encephalopathy that highlights the differences in the effects of encephalopathy in developing and developed populations where mild encephalopathy has not been associated with death or neurodevelopmental disability.26 In the Kathmandu study, moderate encephalopathy was associated with a 71% risk of severe impairment or death, whereas severe encephalopathy had a 97% risk of death or severe impairment, outcomes clearly much worse than those in our population.
The Griffiths Mental Development Scales were selected as the assessment tool for use in our study because it is the developmental instrument used most widely in Australia, and we anticipated that, in the event that we could not test the child ourselves, we would be able to use the results of testing by another pediatrician or psychologist.27 In addition, the Griffiths Mental Development Scales is an age-adjusted assessment with the 5 subscales of development being equally weighted in difficulty.18 Using an age-adjusted psychometric assessment allows valid comparisons of children at different ages and was therefore an appropriate measure for our cohort of children, who were assessed between the ages of 12 and 24 months.
To date, the limited data available on developmental outcomes after newborn encephalopathy have been inconsistent. The variability in reported findings that used the Griffiths assessments as an outcome measure may reflect the traditional use of a mean GQ of 100 with an SD of 12.2 to determine levels of impairment. It is essential that researchers who use older versions of the Griffiths Mental Development Scales18 remain cognizant of the increase in the general and subscale quotients because the Griffiths was originally standardized in 1950 (0–2 years) and 1960 (extended 0- to 8-year version). In comprehensive studies undertaken in the 1980s, the Griffiths quotients obtained with the original 1950 and 1960 samples were compared with 1980 samples.28,29 These results indicate that, using the extended 0 to 8 years of the Griffiths, the 1980s sample of 217 children had a mean GQ score of 111.7 and an SD of 12.7 compared with the 1960 sample with a GQ of 100.2 and an SD of 12.8. These differences were statistically significant even after adjustments had been made for a slightly skewed social class distribution and regional differences. It is reassuring that our control population mean GQ was 113 (SD: 9.3), which is in keeping with more recent study results that used the older version of the Griffiths Scales.1,5,28–30 Because we had available control data that were from a large random sample of infants born at the same gestation, it seemed more appropriate to use our own control population to define the population mean and determine the cutoff for clinically significant developmental delay.
The patients with a history of severe newborn encephalopathy had a poorer outcome than those with moderate encephalopathy. In part, this is explained by the fact that death in the neonatal period was 1 of the defining features of severe encephalopathy. However, even after excluding the neonatal deaths, twice as many of the patients with severe encephalopathy remained in the poor outcome groups compared with the patients with moderate encephalopathy.
Patients with a history of seizures in the neonatal period had a significantly higher rate of cerebral palsy than those without, although adverse neurodevelopmental outcomes other than cerebral palsy were similar among patients with and without seizures. Recent work by Temple et al8 suggests that seemingly normal survivors of neonatal seizures may have a high incidence of specific learning disabilities and poor social adjustment, which emphasizes the need for careful long-term follow-up of these children.
Although it would have been preferable to be able to present this data broken down by clearly defined causative subgroups, the reality is that in most cases of newborn encephalopathy the exact cause is not known. We previously identified many associations in the preconceptional, antepartum, and intrapartum periods, but the mechanisms of action of these risk factors are currently predominantly unknown and it is by no means certain that all associations are causative. Added to this is the genuine difficulty experienced in clinical practice in adequately defining and diagnosing birth asphyxia and the presence of several subgroups of predisposition among those considered to have asphyxia. Therefore, the clinician is usually left with giving more generic advice to the parents of infants with encephalopathy, and this study seeks to provide a basis for this advice. Future papers will seek to clarify this further by analyzing the data according to the type of risk factors identified and the epoch of pregnancy in which they are likely to be active.
Our findings are important as they provide valuable prognostic information regarding survival and the development of disability. It is important to note that the proportion of children with cerebral palsy is likely to increase, as at the time of this analysis, the youngest children in the study were only 3 years old and the diagnosis of cerebral palsy cannot be regarded as certain until age 5 years. Furthermore, notification to the Western Australian Cerebral Palsy Register is often delayed until the diagnosis is confirmed.
In this follow-up article, we present the only population-based data on developmental outcomes of newborn encephalopathy available in the current literature. Newborn encephalopathy in this study was a clinical diagnosis and was made without assumption or exclusions based on cause.
Our findings indicate the need for continued individual clinical follow-up so that appropriate early interventions and educational provisions can be made or that parents can be reassured that their child is progressing normally and will not require special educational provisions. We are continuing to follow up both the patients and control subjects from our original case-control study. This will allow us to identify the obvious and the more subtle effects that a history of newborn encephalopathy can have on behavior, learning, and cognitive function and to determine their significance as the children grow up.
This study was funded by the Public Health Research and Development Committee of the National Health and Medical Research Council of Australia (94/3368) and the National Health and Medical Research Council of Australia (96/0560, 96/3209, 98/7062, 00/3209).
G.D. assessed the majority of the children using the GMDS, conducted some of the analysis, wrote the first draft of the paper, and helped plan the follow-up. N.B. conducted the original case-control study; planned, instigated, and supervised the follow-up; and contributed to the analysis and redrafting of the paper. J.J.K. contributed to the original case-control study, supervised and helped plan the follow-up, conducted most of the analysis, and contributed to the redrafting of the paper. J.M.K. contributed to the original case-control study and contributed to the redrafting of the paper. S.S. helped plan the follow-up, provided clinical supervision for the follow-up assessments, and contributed to the editing of the paper. S.R.Z. helped plan the follow-up, provided clinical supervision for the follow-up assessments, and contributed to the editing of the paper. F.J.S. conceived the original case-control study, helped plan the follow-up, and contributed to the editing of the paper.
We thank the children and parents who participated in this study and the staff of the Telethon Institute for Child Health Research and the Princess Margaret Hospital for Children. We are also grateful to the contributions of Jane Doyle, Fiona O’Sullivan, Margaret Baron-Hay, Patrick J. Pemberton, Linda Watson, Sharon Vukovich, Gail Reading, and Peter Cosgrove.
- ↵Nelson KB, Ellenberg JH. Neonatal signs as predictors of cerebral palsy. Pediatrics.1979;64 :225– 232
- Scher Ms, Aso K, Beggarly ME, Hamid MY, Steppe DA, Painter MJ. Electrographic seizures in preterm and full-term neonates: clinical, associated brain lesions, and risk for neurologic sequelae. Pediatrics.1993;91 :128– 134
- ↵Legido A, Clancy RR, Bergman PH. Neurologic outcome after electroencephalographically proven neonatal seizures. Pediatrics.1991;88 :583– 596
- ↵Badawi N, Kurinczuk JJ, Keogh JM, et al. Antepartum risk factors for newborn encephalopathy: the Western Australian case control study. BMJ.1998;317 :1549– 1553
- ↵Badawi N, Kurinczuk JJ, Keogh JM, et al. Intrapartum risk factors for newborn encephalopathy: the Western Australian case control study. BMJ.1998;317 :1554– 1558
- ↵Griffiths R. The Abilities of Young Children. A Comprehensive System of Mental Measurement for the First Eight Years of Life. London, England: Child Development Research Centre;1970
- ↵Clayton D, Hills M. Statistical Models in Epidemiology. Oxford, England: Oxford University Press;1993
- ↵Reynell J. Manual for the Reynell-Zinkin Scale. London, England: National Foundation of Education Research;1979
- ↵Bayley N. Bayley Scales of Infant Development. 2nd ed. New York, NY: The Psychological Corporation;1993
- ↵Thorndike RL, Hagen EP, Sattler JM. Stanford-Binet Intelligence Scale. 4th ed. Chicago, IL: DLM Teaching Resources;1986
- ↵Herst J, Wolfe S, Jorgensen G, Pellan S, Neel Pefley D. Sewall Early Education Developmental Program: SEED Profile of Development. Denver, CO: Sewall Centre of Rehabilitation;1976
- ↵Peliowski A, Finer NN. Birth asphyxia in the term infant. In: Sinclair JC, Bracken MB, eds. Effective Care of the Newborn Infant. Oxford, England: Oxford University Press;1992:249– 279
- Copyright © 2002 by the American Academy of Pediatrics