Objective. It is not known when behavior problems begin in children with epilepsy. The purposes of this study were to: 1) describe the rates of behavior problems in children before their first recognized seizure, 2) determine the differences in behavior problems between children with a first recognized seizure and their healthy siblings, and 3) identify the seizure variables early in the course of the condition that are associated with behavior problems before the first recognized seizure.
Methods. The sample was 224 children (4–14 years old) with a first recognized seizure and their 135 healthy siblings. As part of a larger study, computer-assisted structured telephone interviews were conducted with mothers to measure child and sibling behavior problems. Behavior problems were measured using the Child Behavior Checklist. Frequencies, t tests, correlational analysis, and multiple regression were used to analyze data.
Results. Higher than expected rates of behavior problems in the 6 months before the first recognized seizure were found in the total seizure sample, with 32.1% being in the clinical or at-risk range. Rates were highest in children who had previous events that were probably seizures, with 39.5% in the clinical or at-risk range. Children with seizures had significantly higher Total, Internalizing, Attention, Thought, and Somatic Complaints problem scores than their nearest-in-age healthy siblings. Within the seizure sample, variables significantly associated with behavior problems after adjusting for research site, child sex, child age, and socioeconomic status (as represented by primary caregiver's education) were interactions of previously unrecognized seizures with gender and epilepsy syndrome/type of seizures.
Conclusions. Children with previously unrecognized seizures are already at increased risk for behavior problems at the time of their first recognized seizure. These findings are consistent with the hypothesis that in some children, epilepsy is a pervasive condition that includes both seizures and behavioral problems.
Past epidemiologic studies have consistently demonstrated that children with epilepsy have higher rates of behavior problems than children with other chronic physical conditions.1,2 It is not known when behavior problems begin in children with epilepsy, though it has usually been assumed that they begin after the diagnosis of epilepsy. Factors accounting for these behavior problems also have not been well delineated.3 Presumed causes include effects of seizures, effects of antiepilepsy medication, poor child and family adaptation to seizures, and neurologic dysfunction that brings about both seizures and behavior problems.4,5 With few exceptions, past studies investigating behavior problems have been conducted on children who had been diagnosed with epilepsy for many years, which made it difficult to separate preexisting causes of behavior problems from those related to living with chronic epilepsy. In a case study of a 4½-year-old boy, however, hyperactivity was found to precede by 18 months the onset of benign partial epilepsy.6
Five studies describing behavior problems in children with new-onset seizure conditions were found. Only 2 of these, however, had investigation of behavior problems as the primary purpose. The first was a large study by Hoare5,7 that compared psychiatric disturbance across several small groups (n = 23–36) of children: new-onset epilepsy, chronic epilepsy, new-onset diabetes, chronic diabetes, new-onset epilepsy controls, new-onset diabetes controls, chronic epilepsy controls, chronic diabetes controls, new-onset epilepsy siblings, and chronic epilepsy siblings. Children with new-onset epilepsy (n = 29) were diagnosed and treatment was initiated within the 3-month period before data collection. Psychiatric disturbance was based on parent report on a well-established behavior questionnaire.8 The highest rates of problems were found in children with epilepsy, with 45% of those in the new-onset and 48% of those in the chronic epilepsy group showing evidence of psychiatric disturbance. In contrast, only 17% of the children with new-onset diabetes and 10% of the controls for the new-onset epilepsy groups showed psychiatric disturbance.5Moreover, fewer siblings of children with new-onset epilepsy were found to have psychiatric disturbance (9%) than siblings of children with chronic epilepsy (30%), leading Hoare7 to conclude that epilepsy also might lead to mental health problems in siblings. A strength of this study was the use of multiple comparison samples. A major limitation, however, was the measurement of behavior up to 3 months after initiation of treatment in the new-onset epilepsy sample, which made it difficult to separate psychiatric disturbance from effects of antiepileptic medication or early response to the diagnosis of epilepsy. Moreover, information was not reported on how long the children in the new-onset epilepsy group had been having seizures before their diagnosis and treatment.
In the second study,4 data were collected within 6 weeks of a first recognized seizure (baseline) and parents were asked to rate their child's behavior in the 6 months previous to their child's first seizure using the Child Behavior Checklist (CBCL).9Results from this study indicated that 24% of the children either had behavior problems in the clinical range or were at risk for such (T score ≥60). Follow-up data were collected 4 months after the first seizure. At follow-up, 30 of the 42 children were considered to have new-onset epilepsy because they had had additional seizures and were placed on medication. The remaining 12 children had no additional seizures. When the baseline scores of the children in the 2 groups were compared, the children with new-onset epilepsy had higher mean total behavior problems scores (M = 55.5) than the children who had no further seizures (M = 48.7). Major limitations of this study, however, were the small sample size and the failure to obtain information about possible previous unrecognized seizures. Nevertheless, these results suggested that children who go on to develop epilepsy might be more at risk for behavior problems even at the time of seizure onset. The authors hypothesized that central nervous system dysfunction might be a causal factor in both seizures and behavior problems.
Information on behavior problems in children with new-onset epilepsy also was found in 3 other studies. In 2 of these studies,10,11 the major purpose was to investigate the effects of initiating antiepilepsy medication in children with new-onset epilepsy. Specific information on the percentages of children with problems before treatment was not provided for either of these studies. Stores and colleagues10 specifically noted that some of the 63 children with epilepsy in their study had psychological abnormalities before treatment and hypothesized that these problems were a result of the epileptic process. In the study by Williams and colleagues,11 parents rated their children's behavior on the day they were diagnosed with epilepsy using the CBCL.9The 37 children with epilepsy had a mean baseline score on Internalizing Behavior Problems that was 8.3 points higher than population norms (M = 58.3). The mean Externalizing Behavior Problems score, however, was slightly below the population norm (M = 49.3). A strength of these studies was the measurement of behavior before initiation of antiepilepsy medication. Neither study, however, provided information on the length of time the children had had seizures before diagnosis nor the exact period over which parents were instructed to rate their children's behavior. Thus it was not possible to separate any effects of having recent seizures from preexisting behavior problems in either of these studies. The third study12 was a series of case studies on 15 children with epilepsy that included parent reports of behavior problems before seizure onset. In 11 of the 15 children, parents reported that behavior problems existed previous to seizure onset.
All of the above studies were limited by small samples and only the study by Dunn and colleagues4 specifically had the parents rate their child's behavior before the first recognized seizure using a well-validated scale. Nevertheless, the high rates of problems found in these studies suggested that a systematic investigation of behavior problems in children before their first recognized seizure was needed to provide important information about when behavior problems begin, as well as possible causes of these problems. If high rates of behavior problems before the first recognized seizure were found, then these problems cannot be attributed to side effects of medication or to a negative emotional response to the diagnosis of epilepsy, such as concerns about stigma. Moreover, comparing behavior problems in children with a first seizure to their nearest-in-age healthy sibling should help to control for the influence of family environmental variables as a possible cause of behavior problems. Finally, to investigate the possible effects of the characteristics of the seizure condition on behavior problems, data were collected in this study on any unrecognized past seizures. The inclusion of children with unrecognized previous seizures provides an unprecedented opportunity to study behavior problems in a sample of children with untreated seizures. Specific research questions were:
What are the prevalence rates of behavioral problems in children with seizures and their healthy siblings during the 6-month period before their first recognized seizure?
What are the differences in behavior problems between the children with seizures and their healthy siblings?
Within the seizure sample, which seizure variables (previous unrecognized seizures, epilepsy syndrome/seizure type) account for the variability in behavior problems after adjusting for research site, child's sex, child's age, and socioeconomic status as measured by primary caregiver's education level?
Children with a first recognized seizure, who were 4 through 14 years old, were prospectively recruited through electroencephalogram (EEG) laboratories, emergency rooms, and pediatric neurologists at 2 large children's hospitals (Indianapolis and Memphis) and from practices of private pediatric neurologists in Indianapolis. Exclusion criteria were: developmental milestones below the normal range, placement in classes for the mentally handicapped, another chronic physical disorder, or seizures precipitated by an acute transient event, such as intracranial infection, metabolic derangement, or immediate effect of head injury. In addition, children who either had >1 febrile seizure or received treatment for febrile seizures were excluded. Efforts were made to recruit participants within 6 weeks of the first recognized seizure. The mean length of time between the first recognized seizure and recruitment was 35 days. The nearest-in-age healthy sibling of the child with the first recognized seizure served as a comparison sample. Consent was obtained before any data collection.
The 116 girls and 108 boys in the seizure sample had a mean age of 8.4 (standard deviation [SD] = 3.0) years. The sibling sample consisted of 135 children (67 girls and 68 boys) with a mean age of 9.9 years (SD = 3.6; range: 3–19 years). The primary caregiver's highest year of education was used in the analysis to reflect socioeconomic status. The mean number of years of education for primary caregivers was 13.8 years (SD = 2.6). Most were either white (75.4%) or black (21.9%). A higher proportion of participants at the Memphis site (43.0%) were black than at the Indianapolis site (8.7%). See Table 1 for sample information. Data from the primary caregiver were used in the study. In all but 3 children the primary caregiver was the mother.
Seizure variables (ie, seizure type and epilepsy syndrome) were determined by pediatric neurologists (D.W.D. and D.F.R.) based on data from the parent's description of the seizure(s) and the child's clinical records, including reports of clinical examinations by the physician, results of EEGs, and neuroimaging tests (when available). Both seizure type and epileptic syndrome were classified using International League Against Epilepsy criteria.13,14Seizure type was defined using the description of the seizure and by the results of the EEG. Syndrome was determined from the description of the seizure, child's developmental history, results of the neurologic examination, EEG results, and the findings on neuroimaging.
Parents also were systematically interviewed to determine whether their child had had any unrecognized previous seizures before the one that prompted their entry into the study. Parents of approximately one-third (n = 76) of the children reported that their child had previous events that were most likely seizures. In most cases, the family did not know the episode was a seizure and did not seek treatment. In other cases, the family sought evaluation after the episode, but treating physicians believed the event was unlikely to have been a seizure and did not order additional assessment, such as an EEG. A determination of previous seizure was made by D.W.D. based on review of the description of the past and current episodes. Information on seizure variables also is presented in Table 2. The proportion of children having unrecognized previous seizures in this study (one-third) is similar to the proportion found by Shinnar and colleagues15 in a sample of children with first seizures.
Behavior problems were measured by the major caregiving parent's ratings of the child and the sibling on the CBCL.9 The CBCL consists of 118 behavior problem items on which parents rate their child's behavior using 3-point scales of 0 (not true), 1 (somewhat or sometimes true), and 2 (very true or often true). To reduce the possibility that parents might rate seizure activity as behavior problems, parents were instructed during the structured interview to not include any behaviors that might be seizures or related to seizures. Past research has shown the scale to have strong reliability and validity as well as standardized scores normal for age and gender.9 The CBCL provides a Total Behavior Problems score, 2 second-order factor scores (Internalizing Problems and Externalizing Problems), and 8 syndrome scores (Aggressive Behavior, Anxious/Depressed, Attention Problems, Delinquent Behavior, Social Problems, Thought Problems, Withdrawn, and Somatic Complaints).
To calculate the prevalence of behavioral problems in participants with seizures and their healthy siblings, the percentage of children who had scores in the clinical and at-risk ranges were computed for the Total Behavior Problems, for the 2 second-order factors (Internalizing Problems and Externalizing Problems), and for each of the 8 syndrome problem dimensions(Aggressive Behavior, Anxious/Depressed, Attention Problems, Delinquent Behavior, Social Problems, Thought Problems, Withdrawn, and Somatic Complaints). The clinical and at-risk ranges were based on recommendations set by Achenbach.9 For the second-order factors, participants with T scores ≥60 were considered at risk and participants with T scores of >63 were considered in the clinical range. For the 8 syndrome dimensions, participants with T scores ≥67 were considered at risk and participants with T scores of >70 were considered in the clinical range. In addition, because one-third of the seizure sample had previously unrecognized seizures, the percentage of seizure participants and siblings in the clinical or at-risk range was further evaluated within previous-seizure subgroups.
To test for differences between the seizure participants and healthy siblings for each of the dimensions, 2 different statistical methods were used. First, to use all available data, including those from children with no sibling data, independence among all measurements was assumed and 2-sample t tests were used to test for differences between all children with seizures and all healthy siblings. Second, because correlations between scores of seizure participants and their respective siblings were high (.60 <r < .70), differences between seizure participants and healthy siblings also were compared using paired t tests on only the complete sibling pairs. In addition, within the sample of seizure participants, 2-sample t tests were conducted for each dimension to test for differences between seizure participants with previous seizures and those without. Similarly, within the sibling sample, 2-sample t tests were conducted to test for differences between siblings of participants with and without previous seizures.
To determine the relationship of seizure variables to behavior problems for the seizure sample, a multiple regression analysis was used to model the Total Behavior Problems score as a function of seizure variables (previous seizures, epilepsy syndrome/seizure type), adjusting for demographic variables (research site, child's sex, child's age, and primary caregiver's education). Adjusted means16 were used to estimate the effects after adjusting for other variables. Model selection was done in 3 stages. In the first stage, investigators J.K.A. and D.W.D. identified the variables of interest. In the second stage, an automatic stepwise selection procedure was utilized to screen out the variables that did not help in explaining the variation of the Total Behavior Problems score (Entry criterion: P < .5, Exit criterion: P> .5). In the third stage, a backward model selection procedure including all the variables that were significant at P= .1 was used. If an interaction was included, then the main effects that comprised that interaction also were included.
As this is an observational study, we have focused on descriptions of the samples and on basic comparisons between groups. Hypothesis testing was used to describe the sample and we have not made adjustments for multiple comparisons.
The percentages of children in the seizure and healthy sibling samples with scores in the clinical range for each behavior problem dimension are presented in Table 3. Also, the percentages of children in the clinical range (T score >63) in the subgroups based on previous unrecognized seizures are presented inTable 3. Total Behavior Problems scores in the clinical or at-risk range (T score ≥60) for the seizure sample was 32.1% and for the sibling sample was 23.0%. For the group of participants with previous seizures, the percentage of participants with Total Behavior Problems scores in the clinical or at-risk range was 39.5%, while the percentage at risk in the subgroup with no previous seizures and in each of the sibling subgroups (siblings of those with and without previous seizures) was lower, ranging from 21.4% to 28.4%. Similarly, the percentage of participants with Total Behavior Problems scores of greater than 63 at baseline was 34.2% for the previous seizure group and ranged from 15.7% to 21.7% for the other subgroups. For comparison purposes, ∼10% of children in the general population should be in the clinical range on the Total, Internalizing, and Externalizing Behavior Problems scores and ∼2% would be in the clinical range on the 8 syndrome scores.9
Means and standard deviations for the behavior problems scores for the seizure sample and for the healthy sibling sample are presented inTable 4 for the total samples. The mean Total Behavior Problems T score was 55.9 for children with seizures, which is approximately 0.6 SD above the population norm on the CBCL.9 Mean scores for Internalizing Problems and Externalizing Problems were lower at 53.4 and 53.8, respectively. Scores on the 8 syndrome dimensions for the seizure sample ranged from 54.9 for Somatic Complaints to 58.5 for Attention Problems. In comparing the seizure participants to their healthy siblings, the results of the 2-sample t tests and of the pairedt tests were similar and, therefore, only the results of the 2-sample t tests are presented in Table 4. Children with seizures had significantly higher Total Behavior Problems and Internalizing Problems scores than their siblings. Analysis of the 8 syndrome scores indicated higher Attention Problems, Thought Problems, and Somatic Complaints scores for the children with seizures compared with their siblings. Within the seizure sample, participants with previous seizures had significantly higher Total Behavior Problems, Internalizing Problems, Anxious/Depressed, Attention Problems, and Thought Problems scores than those without previous seizures (Table 5). The mean for Total Behavior Problems was 57.9 for the children with previous seizures and 54.8 for those without previous seizures. For the sibling sample, there were no differences between the 2 sibling subgroups for any of the behavior variables. The mean Total Behavior Problems score was 53.6 for siblings of children with previous seizures and 51.8 for siblings of children without previous seizures.
A multiple regression analysis was used to model the Total Behavior Problems score as a function of seizure characteristics (previous seizures and epilepsy syndrome/seizure type), adjusting for demographic variables (research site, child's sex, child's age, and primary caregiver's education). Total number of observations included in the regression analysis was n = 223 (attributable to missing values for 1 of the children). To reduce the number of variables in the final model, those that were not statistically significant at 0.10 were dropped from the model. Variables included in the final regression model were: previous seizures (presence/absence), epilepsy syndrome/seizure type, child's age, child's sex, site, primary caregiver's education, and the interactions of sex-by-previous seizure and epilepsy syndrome/seizure type-by-previous seizure. The final model (Table 6) was significant (adjusted R2 = .118, P< .0025). There was a marginally significant effect for site, with children in Memphis having on average more problems than children in Indianapolis (adjusted means: 58.0 vs 55.5). Primary caregiver's education also was significantly related to behavior problems, with lower scores for educational level corresponding to higher scores on behavior problems. After adjusting for demographic factors, there were significant effects of interactions of previously unrecognized seizures with child's sex and epilepsy syndrome/seizure type. The adjusted means corresponding to these interaction effects are presented in Table 7. Boys with previous seizures had the most problems. The adjusted means for boys with and without previous seizures were 60.0 and 54.7, respectively, and for girls with and without previous seizures were 55.5 and 56.8, respectively. (SeeTable 8 for means of behavior problem scores by child's sex and previous seizure status.) In addition, children with partial seizures (included in the other category in the table) and previous seizures showed higher rates of behavior problems than children with primary generalized epilepsy syndromes (absence or tonic-clonic seizure types) regardless of previous seizures (see Table 7).
In this study of behavior problems in children at the time of their first recognized seizure there were 3 major findings. First, there were higher than expected rates of children with seizures who had behavior problems in the at-risk or clinical range before their first recognized seizure. Second, compared with their nearest-in-age healthy siblings, children with seizures had more behavior problems in several areas. Finally, seizure variables at baseline were significantly associated with behavior problems after adjusting for demographic variables. Each of these major findings is discussed.
The finding of increased rates of behavior problems in children before their first recognized seizure is consistent with our previous pilot study finding.4 In this study approximately one-third of the children had a total behavior problems score in the at-risk or clinical range. Although the rate is higher than that found in our pilot study (24%), it is lower than the rate of psychiatric disturbance found in children with new-onset epilepsy (45%) by Hoare.5,7 In addition, the mean internalizing problems score for the children with seizures in this study was approximately one half standard deviation lower than the baseline internalizing problems score reported by Williams and colleagues.11 When the 8 syndrome areas were explored in this study, the area with the highest percentage of children in the clinical range was attention problems. Higher rates of attention problems have been previously reported in children with epilepsy. Semrud-Clikeman and Wical17 found attention problems in children with complex partial seizures. In addition, Williams and colleagues18found children with epilepsy to have subtle attention difficulties.
Children with seizures had higher scores than their siblings in total, internalizing, attention, thought, and somatic complaints problem areas. Mean differences between siblings were relatively small, however, and ranged from 2.3 to 3.6 points. The greatest differences were found in the areas of total behavior and attention problems. Interestingly, with a few exceptions, both sibling samples had mean behavior problems scores that were above the norms. In the only previous study that included data on behavior problems in siblings of children with seizures, Hoare7 found increased rates of behavior problems in the siblings of children with chronic epilepsy (45%) but not in the siblings of children with newly diagnosed epilepsy (9%).
One possible explanation for the increased rates of problems in siblings found in this study is that parents' ratings were inflated because of the stress engendered by the recent onset of seizures in one of their children. Another potential reason might be that behavior problems previously being exhibited in the children with seizures had had a negative effect on the family environment, parenting behaviors, or the behavior of their siblings. Future research on siblings should be conducted to determine whether siblings also are at risk for behavior problems.
A major finding is the diversity related to behavioral functioning within the sample of children who presented clinically with a first recognized seizure. The key risk factor for increased behavior problems was the presence of previously unrecognized seizures. To our knowledge we are the first to investigate behavior problems in children with unrecognized seizures before their first recognized seizures. Interestingly, the percentage of children with previously unrecognized seizures who had behavioral problems in the clinical range (34.2%) is similar to the rate of psychiatric disturbances found in children with a brain disorder (including seizures) in the Isle of Wight study (34.3%).2
This finding of higher rates of behavior problems at seizure onset in children who were having unrecognized seizures is consistent with Aicardi's19 proposal that epilepsy can be a pervasive condition in which the manifestations include seizures as well as behavioral disturbances. One may reasonably ask what the nature of the mechanism is that might be related to both seizures and behavioral disturbances. Aicardi19 proposes that transient cognitive impairment (TCI) caused by subclinical epileptiform discharges might be this mechanism because TCI can lead to disorganization and suppression of normal brain functioning. One could argue that children with previous seizures would be more likely to have subclinical epileptiform discharges than children without previous seizures.
Support for this hypothesis includes the empirical evidence that epileptiform discharges that are not accompanied by clinically observable seizures may result in brief episodes of impaired cognitive functioning and reduced psychosocial functioning. In a pilot study of 10 children with TCI, Marston et al20 administered antiepilepsy treatment to reduce subclinical discharges with the goal of improving psychosocial functioning. In 8 of the 10 children, psychosocial functioning was improved in conjunction with reduced interictal discharges on the EEG. Binnie21 hypothesizes that episodes of TCI can adversely affect interpersonal interactions or social functioning if it should cause the child to miss important social or emotional cues during interactions with peers, or whether the child experienced interruptions in the flow of conversations, leading to failure to respond appropriately. The increased rates of attention problems found in this study also indirectly lend support for the hypothesis that transient cognitive impairment could account for increased behavioral problems found in children with epilepsy. Limitations of the study by Marston and colleagues,20however, were the small sample size and the global measurement of psychosocial functioning. In contrast, epileptiform discharges were not found to be related to behavior problems in children from the general population.22
Other variables significantly predicting behavior problems before a first recognized seizure were interactions of previously unrecognized seizures with gender and epilepsy syndrome/type of seizure. The finding that male gender was a risk factor for behavior problems was not unexpected because of the previous epidemiologic research showing that during preadolescence boys are more at risk for problems than are girls (eg, Offord et al).23 Reasons for the interaction of male gender and previous seizures, however, are not immediately obvious.
To our knowledge, the exploration of the interactions of epilepsy syndrome/type of seizure with previous seizures has not been previously investigated in children with new-onset seizures. Findings from other studies of children with new-onset epilepsy, however, suggest that behavior problems are higher than expected in children with partial epilepsy syndromes. For example, in a study of children with new-onset temporal lobe epilepsy (ie, secondary partial epilepsy syndrome) 38% were found to have psychological problems.24 In addition, a study by Weglage et al25 showed that children with primary partial epilepsy syndrome had more behavior problems, lower IQ, and poorer neuropsychologic functioning than controls.
Because this was the first large-sample study to systematically investigate behavior problems in children presenting with first recognized seizures, it is imperative that these findings be validated, especially the finding of higher rates of problems in children with previously unrecognized seizures. Although only parent ratings of behavior problems are reported here, results from teacher ratings of behavior problems in our larger study also show higher rates of problems at baseline in children with previously unrecognized seizures. Furthermore, when the higher rates of behavior problems found in this study are examined in conjunction with the increased rates of behavior problems in children with newly diagnosed epilepsy found by others,5,7,10,11 there is substantial empirical support that behavior problems have their onset very early in the course of the disorder.
A major limitation of the current study was that minimal information was collected on the nature of the previously unrecognized seizure(s). Certainly, the strength of the relationship between previous seizure activity and behavior problems strongly suggests that additional information on any previous seizures should be collected in future studies. A complete description of all previous seizure-like episodes would be important to obtain to determine whether the child had had any other types of seizures that the family would not recognize as being seizures. For example, a previous partial seizure may not be even mentioned by a parent who only recognizes a tonic- clonic seizure. In addition to estimating the number and type of previous seizures, it would be important to measure the length of time since the first seizure because the effect of subclinical discharges has been proposed to have a cumulative effect.25
Another limitation was that neuropsychologic functioning (eg, attention, memory, and executive functioning) was not measured in this study. Because TCI is proposed to affect neuropsychologic functioning, which in turn leads to behavior problems, careful attention to neuropsychologic functioning should be part of future studies. Even with these limitations, however, results from this study suggest that children presenting with first recognized seizures should be considered at risk for and assessed for behavior problems. It is especially important to assess for previously unrecognized seizures because these children are particularly at risk for problems. Substantiation of these findings in future research would have important implications for treatment of seizures in children.
This study provides important information about potential causes of behavior problems in children with epilepsy. When these findings are examined in relation to the 4 presumed causes of behavior problems (effects of seizures, effects of antiepilepsy medication, poor child and family adaptation to seizures, and neurologic dysfunction that brings about both the seizures and the behavior problems), the strongest support is found for effects of seizures and/or neurologic dysfunction. Having parents rate their children's behavior problems before the first recognized seizure substantially reduces the possibility that side effects of medication or a negative family response (eg, stigma) to a first seizure accounts for the higher rates of behavior problems.
This research was supported by Grant PHS R01 NS22416 from the National Institute of Neurological Disorders and Stroke to Dr. Austin.
We acknowledge assistance from Drs. B. Hale, B. Garg, H. Patel, J. Pappas, and O. Markand as well as the Epilepsy and Pediatric Neurology Clinics at Riley Hospital, Indiana University Medical Center and the Medical Research Committee of Methodist Hospital, Indianapolis, Indiana. We also thank Angela McNelis for help with data collection and Phyllis Dexter and Jeanne Critchfield for editorial assistance.
- Received November 30, 1999.
- Accepted July 26, 2000.
Reprint requests to (J.K.A.) Indiana University School of Nursing, 1111 Middle Dr, NU403, Indianapolis, IN 46202-5107. E-mail:
- CBCL =
- Child Behavior Checklist •
- EEG =
- electroencephalogram •
- SD =
- standard deviation •
- TCI =
- transient cognitive impairment
- ↵Rutter M, Graham P, Yule W. A neuropsychiatric study in childhood.Clin Dev Med. 1970;35,36
- Rutter M, Tizard J, Whitmore K. Education, Health, and Behaviour. Huntington, NY: Robert E. Krieber Publishing Company; 1970
- Achenbach TM. Manual for the Child Behavior Checklist/4–18 and 1991 Profile. Burlington, VT: University of Vermont Department of Psychiatry; 1991
- Stores G,
- Williams PL,
- Styles E,
- Zaiwalla Z
- Nolte R,
- Wolff M
- Shinnar S,
- Berg AT,
- Moshe SL,
- et al.
- Neter J, Wasserman W, Kutner MH. Applied Linear Statistical Models. 2nd ed. Irwin, Burr Ridge, IL; 1990
- Aicardi J
- Harvey AS,
- Berkovic SF,
- Wrennall JA,
- Hopkins IJ
- Copyright © 2001 American Academy of Pediatrics