 |
|
 |
|
 |
 |
|
 |
 |
 |
 |
 |
 |
 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
,||,||
* Department of Biological Sciences, Northern Illinois University, DeKalb, Illinois
Department of Pediatrics, Yale Medical School, New Haven, Connecticut
Department of Mathematics, Northern Illinois University, DeKalb, Illinois
|| Department of Neurology, Yale Medical School, New Haven, Connecticut
¶ Departments of Neurology and Pediatrics, Montefiore Medical Center and Albert Einstein College of Medicine, New York, New York
 |
ABSTRACT |
|---|
 TOP ABSTRACT METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Methods. As part of a prospective community-based study of newly diagnosed epilepsy, parents of children who were younger than 3 years at the time of initial onset of epilepsy completed the Vineland Adaptive Behavior Scales screener version at entry into the study and once a year thereafter for up to 3 years. Longitudinal analyses were performed on the composite score as the primary outcome and on the 4 domain scores (communication, socialization, motor, and daily living) as confirmatory/secondary outcomes to determine the effects of syndrome (epileptic encephalopathy or other), seizure control (intractable or not), and etiology (symptomatic or not) on adaptive behavior at the time of initial diagnosis and over time.
Results. A total of 613 children were enrolled in the study, and 191 (31%) of these children met the age criterion for this analysis. Of these, 172 (90%) had adequate follow-up and had completed baseline and at least 1 subsequent Vineland assessment. Overall, Vineland scores (composite and individual domains) were somewhat below average at baseline (initial diagnosis). All declined significantly over time. All of the effects at baseline, however, were limited to children with epileptic encephalopathies and symptomatic etiology. Substantial declines over time occurred in these children, and there was an independent effect of intractable seizures as well. In children with none of these factors (
75% of the study group), baseline scores were consistent with average performance for the test norms and there was no evidence of any decline over time.
Conclusions. Children with an underlying symptomatic etiology or a syndrome that can be characterized as 1 of the epileptic encephalopathies demonstrate impaired adaptive behavior at the time of initial diagnosis and experience additional declines in assessments of age-adjusted performance over time. Our results suggest that future seizure outcome is not strongly reflected in adaptive behavior at initial diagnosis but that it takes its toll on the child over time. Understanding how each of these factors affects development and how they interact with each other is the next step in designing effective interventions for lessening the impact of these disorders on the child. In the majority of children with onset of epilepsy during infancy or early childhood, adaptive behavior is within the normal range and does not show any evidence of declining over time. Although this is encouraging, it does not contradict other studies that have demonstrated behavioral and relatively subtle cognitive difficulties in school-aged children with epilepsy. Long-term follow-up in this cohort will permit us to examine the predictive value of the Vineland for later behavioral and cognitive difficulties in this group that, so far, seems to be doing well.
Key Words: epilepsy • adaptive behavior • development • epileptic encephalopathy • longitudinal analysis
Early views maintained that seizures damaged the brain and that children with epilepsy suffered potentially severe consequences as a result of having epilepsy. Ellenberg and Nelson1 provided the first compelling evidence that this was not necessarily the case. In their study, they were able to demonstrate no significant change in full-scale IQ in otherwise neurologically normal children by comparing measures taken before and after the onset of epilepsy. Because of the study design, their analysis considered only children who were older than 4 years at the time of initial onset of epilepsy. The epilepsies are a highly diverse group of disorders with different types of epilepsy occurring at different ages.2 Onset of epilepsy during infancy and very early childhood may be particularly problematic as this is the age when some of the most serious forms of epilepsy, the epileptic encephalopathies, initially occur.3 One population-based study examined the seizure outcomes in children who were younger than 2 years at the onset of the epilepsy and found etiology to be the strongest determinant of both seizure control and normal mental development.4 The role of syndrome type and the independent influences of etiology, seizure control, and syndrome type on development are still not well understood.
In a prospective, community-based cohort, we examined over the first 3 years after the initial diagnosis the performance on the Vineland Scales of Adaptive Behavior of young children who received a diagnosis of epilepsy. Our specific goals in this set of analyses were to determine whether adaptive behavior was consistent with expectation for age at initial onset and whether it changed over time. Furthermore, we examined the independent roles of syndrome type, etiology, and seizure control on changes in the age-appropriate levels of adaptive behavior over time.
|
METHODS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Epilepsy syndrome,6 seizure type(s),7 and etiology8 were classified by 3 pediatric neurologists (S.S., S.R.L., and F.M.T.) according to current, published criteria from the International League Against Epilepsy. Classifications were performed at initial study entry and then again 2 years later. We have used the syndrome and etiology as classified at 2 years. In particular, we identified children whose syndrome fell in the class of syndromes now considered to be epileptic encephalopathies.3 These include but are not limited to well-known and fairly common syndromes such as West and Lennox-Gastaut. In this age group, most children whose syndromes are not within the epileptic encephalopathy cluster have a nonidiopathic localization-related syndrome or an unclassified form of epilepsy.
Because there are very few children with idiopathic syndromes in this age group, we combined idiopathic and cryptogenic and compared this combined group (nonsymptomatic) with children with remote symptomatic causes. Remote symptomatic is used to indicate that an individual has a known brain insult or condition that is associated with an increased risk of epilepsy.8 This evidence may come from the history, examination, neuroimaging study, or other sources. The notion is that the condition or insult is, at least in part, responsible for the occurrence of seizures in that individual. Seizures that occur in the immediate context of an acute insult are considered acute provoked or acute symptomatic and are not part of the spectrum of epilepsy proper (recurrent unprovoked seizures).
Children were followed prospectively from the time of initial diagnosis by telephone every 3 months. Information regarding the occurrence of seizures was collected and summarized by year of follow-up. Seizure outcome was classified as intractable when a child failed trials of 2 antiepileptic drugs and had an average of at least 1 seizure per month over an 18-month period during the first 3 years of follow-up.
The Vineland Scales of Adaptive Behavior Screener was administered to the parent by a research associate at the time of the initial intake interview and once a year thereafter.9,10 The Vineland focuses on behaviors needed to operate in an age-appropriate manner in the individual's environment. It thus is age specific. The Vineland provides an overall composite score as well as domain scores for communication, daily living skills, socialization, and motor skills. For infants and very young children, the behaviors assessed are closely correlated with developmental milestones. The Vineland scores are standardized to a mean of 100 and a standard deviation of 15 based on a normative population sample.9
Children who were followed for <3 years and those who did not have a baseline Vineland or were missing all follow-up Vineland administrations were excluded. Other than limiting analyses to children with reasonably complete data, we did not make additional adjustment or use imputation methods for missing Vineland observations.
Analyses were conducted using t tests and analysis of variance for bivariate comparisons. To examine the longitudinal changes in Vineland scores, we used a longitudinal multiple regression analysis algorithm that takes into account the intercorrelations between repeated measures made in an individual. Analyses were performed to determine the associations of the covariates with the overall composite score and with each of the 4 domain scores. Interaction terms were used to reflect the changes in scores within certain subgroups over time. The algorithm provided by the "Mixed" procedure in SAS was used for longitudinal analyses.11 This procedure provides a flexible means to account for the correlation between the 4 scores measured on each patient for each domain. Schwarz' Bayesian Information Criterion was used to select an appropriate correlation structure for each domain.
|
RESULTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
The distribution for the age at onset (first unprovoked seizure) by 6-month intervals (<6, 6–11, 12–17, 18–23, 24–29, 30–35 months) was 44 (25%), 32 (19%), 24 (14%), 15 (9%), 23 (13%), and 34 (20%). Poor seizure control sufficient to meet our criteria for intractability was evident in 29 (17%) of the children.
Etiology was considered nonsymptomatic (ie, cryptogenic or idiopathic) in 124 (72%) of the children and symptomatic in the other 48. Specific remote symptomatic etiologies included intrauterine insult (n = 13); intraventricular hemorrhage, stroke, or hypoxic encephalopathy (n = 6); brain malformations (n = 11); previous central nervous system infection (n = 6); neurocutaneous syndrome (n = 3); chromosomal anomaly (n = 5); and other (n = 4).
The syndrome was considered 1 of the epileptic encephalopathies in 50 (29%) children. These included 18 West syndrome, 13 Lennox-Gastaut syndrome, 6 Doose syndrome (or myoclonic astatic epilepsy), 2 Dravet syndrome (or severe myoclonic epilepsy of infancy), 8 symptomatic generalized epilepsies, and 3 others. In the remaining 122, the syndromes were classified as symptomatic (n = 23) or cryptogenic (n = 66) localization-related epilepsy, idiopathic generalized epilepsy (n = 8), and unclassified epilepsy (n = 25).
Vineland Scores Initially and Over Time
The Vineland scores at initial entry into the study reflected adaptive behavior competency that was slightly below average. Over time, significant declines in the composite as well as all 4 domain scores were observed (Table 1).
|
|
|
|
12 points (nearly a standard deviation) every year. Significant decreases over time were seen in 2 or 3 of the 4 domain scores for each of these 3 time interactions, with daily living and communication being the most consistently involved. Because the epileptic encephalopathies are typically characterized by failure to acquire normal developmental milestones and sometimes even declines in function, we also performed the longitudinal analysis separately in those with and without these types of syndromes. We found similar patterns of effects in the 2 groups. Specifically, there was an independent negative effect of remote symptomatic etiology at baseline and over time. Children with poor seizure outcomes experienced declines in adaptive behavior scores over time, although they were not significantly worse off at the time of initial diagnosis. The only difference that we found in the 2 groups defined by syndrome type was that there was a small drop in Vineland scores with time in the epileptic encephalopathy group even after adjusting for etiology and seizure outcome (–3.6 points per years; P = .031) for the composite score. Of the 4 domains, this effect was statistically significant only for the communication score (–3.9 points per year; P = .029). These results were largely evident already in the overall longitudinal analysis.
Figure 1 emphasizes the main findings of these analyses. Neurologically intact children whose seizures are not intractable and whose syndrome was not among those in the epileptic encephalopathy cluster initially seem average (normal) on measures of adaptive functioning. Their continued average adaptive function scores over the first 3 years after diagnosis indicate that they acquired the necessary skills to maintain an age-appropriate performance over time. By contrast, children with one of the syndromes in the epileptic encephalopathy cluster, intractable seizures, or remote symptomatic etiology tend to be impaired from the start. The drop in their age-adjusted scores over time indicates that they did not acquire the necessary skills at the necessary rate to maintain even their relatively lower age-adjusted adaptive behavior level.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Our results clarify certain issues. Children with syndromes in the epileptic encephalopathy group do poorly even after controlling for symptomatic etiology and poor seizure control, although some of the effect of these syndromes is partially explained by these other 2 factors. This is not at all surprising and very much conforms to what one would expect on the basis of the detailed clinical descriptions of these syndromes and their prognoses.2 These are syndromes marked by deterioration in broad aspects of neurologic function.3
In children who have remote symptomatic conditions but did not have syndromes classified as one of the epileptic encephalopathies, adaptive function is significantly and substantially lower at initial diagnosis of epilepsy. What was not anticipated was that measures of adaptive function would continue to decline over time in this group. This suggests that these children failed to acquire sufficient age-appropriate skills over time either because of a loss of skills in some or, more likely, a failure to acquire new skills or acquisition of new skills at an inadequate rate. Overall, this is a group that, given its high risk for poor developmental outcomes, may be targeted for aggressive early intervention and, for the most part, already is.
In the majority of the study group—children who have nonsymptomatic epilepsy and whose syndromes were not part of the spectrum of epileptic encephalopathy—there was no evidence, at the time of diagnosis, that these children were in any way disadvantaged with respect to adaptive behavior. Furthermore, over time, there was no evidence that their adaptive functioning declined. This suggests that they developed relatively normally. Only 3 children in this group developed intractable seizures. Despite the small number, there was evidence of a substantial decline in adaptive function in this group. For example, the composite score fell from 84 at initial study entry to 54 at 3 years. It would be preferable to validate this finding in a much larger sample; however, our initial interpretation is that this result suggests an independent effect of uncontrolled seizures on development.
The effects that we found over time (modeled by interactions) were evident in the composite scores and also largely reflected in the individual domain scores. All domain scores showed declines over time in children with any of the 3 main features that we considered (etiology, epileptic encephalopathy, and intractability). Statistically significant declines were most consistently seen for the daily living skills domain and least consistently for socialization. Analysis of individual domain scores may or may not shed additional light on the effects that epilepsy has on these children; however, the extent to which the results are consistent helps to bolster interpretation of findings for the composite measure.
Because there are many facets of behavior that can be measured and studied, it is essential to recognize what the Vineland Scales do and do not reflect. They measure adaptive behaviors that an individual, in this case an infant or very young child, routinely performs in his or her everyday life. Information is collected by interviewing the parent. The Vineland focuses on very basic routine behaviors that depend on social, communicative, and motor skills. For this reason, it is a good reflection of development, as perceived and reported by the parent(s), during the first several years of life. Because of its brief, standardized administration, it is also suitable for repeated use in a large research study such as this. Adaptive behavior is not, however, the same as a detailed neuropsychological examination in an older child or an adult. It also is not a measure of behavioral, cognitive, and psychiatric disorders, many of which are not readily diagnosable until later in childhood and adolescence. Consequently, we could not address whether these children, when they go to school, will perform as expected or have selective difficulties, behavioral disorders, or any of a host of other problems that are simply not reflected in what the Vineland was designed to measure. For these reasons, our results do not contradict in any way the findings of other groups that have demonstrated a variety of difficulties in children with epilepsy.12,13 Evidence of high levels of psychiatric and behavioral disorders and educational, neuropsychological, and social problems have been repeatedly reported in children with epilepsy.14 Population-based studies have demonstrated a higher-than-expected likelihood of a variety of negative educational and social outcomes in neurologically normal individuals who, as adults, were completely seizure-free.15,16
All of the findings suggesting poorer-than-expected social and educational outcomes in individuals with childhood-onset epilepsy despite good seizure outcomes are real. These are factors that often cannot be assessed reliably until an individual is much older. The extent to which adaptive behavior measured early in life reflects normal functioning in different areas later in life is largely unknown. In the current phase of this study, we are in the process of inviting participants to have a neuropsychological assessment at 9 years after their initial diagnosis of epilepsy. Comparison of these more detailed neuropsychological results with the measures of adaptive behavior taken years earlier will enable us to determine the extent to which the early measures are able to reflect future performance. For the time being, however, our findings provide reassurance regarding basic development and adaptive behavior in the majority of infants and young children with epilepsy.
|
ACKNOWLEDGMENTS |
|---|
We are especially grateful to all of the physicians in Connecticut who have referred their patients to this study. Eugene Shapiro has kindly facilitated many administrative aspects of the study for us. This study would not be possible without the generous participation of the many parents and their children in this study.
|
FOOTNOTES |
|---|
Reprint requests to (A.T.B.) Neuroepidemiology Group, Department of Biological Sciences, Northern Illinois University, DeKalb, IL 60115. E-mail: atberg{at}niu.edu
|
REFERENCES |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  M. Freilinger, B. Reisel, E. Reiter, M. Zelenko, E. Hauser, and R. Seidl Behavioral and Emotional Problems in Children With Epilepsy J Child Neurol, November 1, 2006; 21(11): 939 - 945. [Abstract] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
||
|
||
Home | My Pediatrics | Journal Information | Current Issue | Past Issues | Subscriptions & Services | Contact Us Click here for faster international access © 2004 American Academy of Pediatrics. All rights reserved. |
||
|
||
