SEARCH

SEARCH BY CITATION

Keywords:

  • Cognitive function;
  • Cognitive impairment;
  • Preschool children;
  • Epilepsy

Summary

  1. Top of page
  2. Summary
  3. Purpose
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. References

Purpose: Studies have shown that underlying pathology and early onset of seizures are both significant factors contributing to cognitive impairment in children with epilepsy. However, there are only few studies focusing on cognitive impairment in preschool children with epilepsy. The purpose of this study was to describe the cognitive performance in a population-based cohort of preschool children with epilepsy. The aims of the study were to determine frequency of cognitive impairment, level of cognitive functions, and epilepsy-related factors correlating with cognitive impairment.

Methods: The study group consisted of a population-based cohort (N = 64) of preschool children (3–6 years 11 months) with active epilepsy. Medical data and results from previous psychological evaluations were reviewed retrospectively from the medical records. A logistic regression model was used for the prediction of cognitive impairment.

Key Findings: Prevalence of epilepsy was 3.2 per 1,000 children. Cognitive function was considered to be within normal or borderline range for 50%, mildly retarded for 22%, and moderately to severely retarded for 28%. Cognitive impairment was related to complicated epilepsy, age at onset of epilepsy, abnormal magnetic resonance imaging (MRI), and additional neurologic problems. Age at the onset of seizures was the only significant predictor of cognitive impairment.

Significance: The results concur with those of earlier studies on cognitive impairment in childhood epilepsy. Age at onset of epilepsy is also an important factor for cognitive impairment on young children with epilepsy. The results suggest that cognitive impairment is evident early in the course of epilepsy.

Many factors may contribute to cognitive impairment in childhood epilepsy. Underlying brain abnormalities are associated with impaired cognitive functions in children with epilepsy (Nolan et al., 2003). Children with symptomatic epilepsy are more likely to have mental retardation (Vasconcellos et al., 2001). Instead, children with idiopathic epilepsy are reported to have cognitive performance at the lower end of normal distribution, although differing from healthy controls (e.g., Høie et al., 2005; Cormack et al., 2007). Generally, the proportion of intellectual dysfunction (IQ <80) in children with epilepsy varies between 26% and 57% (Cormack et al., 2007; Berg et al., 2008) and in population-based studies, about 20–40% of children with epilepsy are reported to have mental retardation (Sillanpää, 1992; Sidenvall et al., 1996; Waaler et al., 2000; Camfield & Camfield, 2007). In addition, Sillanpää (1992) indicated other neurocognitive difficulties; about 28% have speech disorders and 23% learning difficulties. It has also been suggested (Keene et al., 2005) that the cognitive problems in epilepsy are likely to be related to a common preexisting underlying pathology influencing both.

Although the underlying etiology is perhaps one of the strongest predictors of cognitive impairment, several studies have demonstrated that early age at onset of seizures seems to be another major risk factor contributing to cognitive impairment (Bourgeois et al., 1983; Battaglia et al., 1999; Bulteau et al., 2000; Smith et al., 2002; Freitag & Tuxhorn, 2005; Mangano et al., 2005; Cormack et al., 2007), even independently of the pathology of epilepsy (Vasconcellos et al., 2001). Both animal and clinical studies demonstrate that early seizures affect brain development and increase the subsequent vulnerability of the mature brain to the effects of seizures (Squier et al., 2003; Holmes, 2004). Children with seizure onset during the first year of life are especially likely to have poorer cognitive outcome (Vanderlinden & Lagae, 2004; Mangano et al., 2005; Cormack et al., 2007).

In addition to etiology or age at onset, longer duration of epilepsy (Elger et al., 2004) or proportion of life with epilepsy (Smith et al., 2002) has been found to be associated with cognitive function. Some (Aldenkamp & Bodde, 2005) have emphasized the association between seizure control or seizure frequency and cognitive function. Furthermore, current antiepileptic medication (AED) has been found to be associated with cognitive function (Bulteau et al., 2000; Berg et al., 2008), but some (Smith et al., 2002) have reported contradictory findings.

Early onset of seizures is also often related to underlying neurologic pathology, that is, to symptomatic epilepsy (Rodin, 1989; Svoboda, 2004). In addition to early onset of seizures, high seizure frequency or intractable epilepsy has been found to coincide with both symptomatic etiology and additional neuroimpairment (Eriksson & Koivikko, 1997; Ramos-Lizana et al., 2009). Therefore, it is not surprising that mental retardation and cognitive and behavioral problems are common among children with complicated epilepsy (Kobayashi et al., 2001; Vasconcellos et al., 2001). In addition, mental retardation or cognitive problems are overrepresented in surgical candidates with severe refractory epilepsy. For example, Freitag and Tuxhorn (2005) reported that >70% of preschool surgical candidates were mentally retarded (IQ <70). Despite the well-recognized fact that early onset seizures may affect cognitive function, there are only few studies (e.g., Battaglia et al., 1999; Chaix et al., 2003; Mangano et al., 2005) on preschool-aged children. In addition, children with additional neuroimpairment (e.g., cerebral palsy) and learning disability/mental retardation (IQ <70) are often excluded from studies (e.g., Rejnö-Habte Selassie et al., 2008).

Purpose

  1. Top of page
  2. Summary
  3. Purpose
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. References

The purpose of this present study is to describe the full range of cognitive performance in preschool children with epilepsy. First, the aim of this study was to describe the characteristics of a population-based cohort of preschool children with epilepsy and to determine the frequency and the level of cognitive impairment in more detail. Secondly, the associations between the neurologic variables (e.g., age at onset, seizure control, AEDs used) and cognitive functions were studied. For descriptive purposes the cohort was divided into two subgroups: children with uncomplicated (n = 26) and complicated (n = 38) epilepsy. Uncomplicated epilepsy refers to idiopathic or cryptogenic epilepsy and complicated epilepsy refers to epilepsies with remote symptomatic cause or an epileptic encephalopathy (Berg et al., 2008). On the basis of earlier studies conducted on school-aged children with epilepsy, it was first hypothesized that the intellectual functioning of children in the uncomplicated group would be within normal range and impaired among children with complicated epilepsy (Mandelbaum & Burack, 1997; Bailet & Turk, 2000; Vasconcellos et al., 2001). The second hypothesis was that cognitive impairment in preschool children would be associated with underlying etiology, age at onset of seizures, duration of epilepsy, and seizure frequency that studies (Battaglia et al., 1999; Vasconcellos et al., 2001; Elger et al., 2004; Cormack et al., 2007) have previously shown to contribute to cognitive functions.

This research was conducted in collaboration with the Department of Psychology at the University of Tampere, the Pediatric Neurology Unit of Tampere University Hospital, and the Pediatric Research Centre at the University of Tampere. The ethical committee of Tampere University Hospital approved the study. Parents of children participating in psychological assessment gave their informed written consent. For retrospective review of medical and psychological data, the approval of the ethical committee was sufficient.

Methods

  1. Top of page
  2. Summary
  3. Purpose
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. References

Participants

A population-based cohort of preschool-aged children with active epilepsy (N = 64) was identified through the medical records of the Pediatric Neurology Unit, Tampere University Hospital, Finland, which is the only center for pediatric neurology services in the hospital district. On the point prevalence day, September 30, 2004, the total population of the hospital district was 464,976, of whom 19,821 were children aged 3–6 years 11 months. The study cohort consisted of all children aged 3–6 years 11 months with epilepsy.

Medical and psychological assessment data

All medical and psychological records, neurophysiologic recordings, and available clinical data were reviewed retrospectively. Medical data including demographic factors, duration of epilepsy (at the time of psychological assessment), seizure and epilepsy type, seizure control during the period of the study, electroencephalography (EEG), imaging of the brain with 1.5T MRI, current AEDs, and most recent AED levels were analyzed retrospectively from children’s medical records. Examiner’s comments and observations during the psychological assessment were also checked. According to the medical reports there were nine children in the complicated group with daily seizures, no seizures during the assessment day or the previous night were reported by parents. Of the total cohort (N = 64), psychological assessment of cognitive abilities were available for 47 children, 14 children were neurologically too impaired for reliable assessment, and 3 children with idiopathic epilepsy did not consent to psychological assessment. The Finnish versions of Wechsler’s Primary and Preschool Scale of Intelligence – Revised, WPPSI-R (Wechsler, 1995) or Wechsler Scale of Intelligence for Children –III, WISC-III (Wechsler, 1999) were used. Thirty-seven were assessed with WPPSI-R and 10 with WISC-III. Five of the nonparticipating children were assessed at the age of 18–28 months with the Bayley Scales of Infant Development, BSID-II (Bayley, 1994). Due to different nature of the Bayley Scales, these results have been excluded from the analysis. The psychological evaluations were administered by a clinical psychologist or neuropsychologist. For those children unable to participate in standardized psychological evaluation, the level of cognitive functioning was determined from medical and psychological records, parental reports, and, when available, observations from daycare. Cognitive functioning was classified as within normal (IQ ≥70), mildly retarded (IQ = 50–69), moderately retarded (35–49), and severely/profoundly retarded (<34).

Statistical analysis

Statistical analyses were done with the Statistical Package for Social Sciences (SPSS, version 16.0; SPSS Inc., Chicago, IL, U.S.A.). Continuous and normally distributed variables were compared with Student’s t-test, and chi-square test was used for categorical variables. Spearman’s rho was used to explore the linkage between epilepsy-related factors and cognitive impairment. P-values <0.05 were considered statistically significant. A logistic regression analysis was used to predict cognitive impairment. Level of cognitive function was dichotomized into normal (IQ ≥70) and impaired (IQ ≤69, including the severe, unassessable participants). Predictors considered included age at onset seizures, duration, etiology (idiopathic/cryptogenic vs. symptomatic), MRI (normal vs. abnormal), seizure frequency (good, partial vs. poor), and AED (AED vs. no AED). First, all the independent variables were tested by univariate logistic analysis. Thereafter, the predictive model was executed using a multivariate logistic analysis.

Due to the severity of cognitive and other neurologic impairment, we were not able to reliably assess the cognitive functions of 14 children with standardized measures. This unassessable group included children with severe psychomotor retardation without specific etiology (n = 6), infantile neuronal ceroid lipofuscinosis (INCL, Hagberg-Santavuori disease, n = 3), Rett syndrome (n = 2), congenital CNS malformations (n = 2), and Charge syndrome and mental retardation (n = 1). In some instances (e.g., in analyzing IQ scores), this unassessable group was excluded in order to ensure that they did not influence the findings. For the logistic regression analysis cognitive function for the unassessable group was classified as impaired.

Results

  1. Top of page
  2. Summary
  3. Purpose
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. References

Medical and clinical data

The total cohort consisted of 64 children with epilepsy; 33 girls and 31 boys participated in the study. The total point prevalence was 3.2 per 1,000 in children aged from 3 to 6 years 11 months. Mean age was 65.16 months (SD = 13.38) and age at onset of epilepsy varied between 0 and 75 months (mean 26.98 months, SD = 21.05). About 55% of children had normal MRI and 45% had abnormal MRI. Twenty-seven children had focal seizures, 31 had generalized seizures, and six were unclassified. Monotherapy was used in 64.1% children, polytherapy in 26.6% children, and 9.4% of children had no AED. Valproate was the most common AED used (n = 43, 67.1%). Seizure control was good in 37%, partial in 16%, and poor in 47% of the participants. There were differences in epilepsy-related variables between the uncomplicated and complicated groups. The children with complicated epilepsy were significantly younger at the onset of seizures (mean: 21.89 vs. 32.42, t62 = 2.43, p < 0.02) and had, therefore, longer duration of epilepsy (mean: 43.05 vs. 30.42, t62 = −2.07, p < 0.05) than those with uncomplicated epilepsy. They also had higher seizure frequency than children with uncomplicated epilepsy (inline image = 11.8, p < 0.05). Detailed characteristics are presented in Table 1.

Table 1.   Clinical and demographic characteristics of the study groups
 All subjectsUncomplicated groupComplicated group
  1. a Age was calculated at the time of psychological assessment.

  2. b t62 = 2.43, p < 0.02.

  3. c t62 = −2.07, p < 0.05.

  4. d inline image = 11.8, p < 0.05.

N642638
Gender
 Female331221
 Male311417
Age (months)a
 Mean (SD)65.16 (13.38)64.85 (11.61)65.37 (14.61)
 Range36–8343–7936–83
Etiology
 Idiopathic 6 6 0
 Symptomatic27 027
 Cryptogenic312011
MRI
 Normal3526 9
 Abnormal29 029
Age at onset of seizures (months)
 Mean (SD)26.98 (21.05)34.42 (20.93)21.89 (19.83)
 Range 0–75 6–72 0–75b
Duration (months)a
 Mean (SD)37.92 (23.99)30.42 (20.31)43.05 (25.20)
 Range 2–83 2–71 2–83c
Epilepsy type
 Focal27 918
 Generalized311615
 Unclassified 6 1 5
Seizure control
 Daily 9 0 9
 Weekly10 2 8
 Monthly 8 5 3
 Yearly11 6 5
 >1 year remission 6 2 4
 >2 year remission2011 9d
Status epilepticus
 No50 (78%)2030
 Yes14 (22%) 6 8
AED
 No AED 6 (9%) 2 4
 Monotherapy41 (64%)2021
 Polytherapy17 (27%) 413

Cognitive function

Cognitive function was considered to be within the normal range (IQ ≥70) for 50% of the participants. However, some of these children (n = 6, 9%) may be considered to have borderline (IQ 70–79) cognitive function. Mild mental retardation was found in 21.9% of the participants and moderate to severe retardation in 28.1% of the cohort. The mean IQ (n = 47) of the cohort was 76.00 (SD = 26.06), which was significantly lower than normative mean (t46 = −6.32, p < 0.001). As demonstrated in Fig. 1, cognitive impairment was more common in children with complicated epilepsy, and cognitive function in children with uncomplicated epilepsy was within normal range. Children with complicated epilepsy had significantly lower overall IQ (91.27 vs. 62.56, t45 = 4.49, p < 0.001) than children with uncomplicated epilepsy. In fact, the level of cognitive functioning in this group was lower than the mean IQ, since the neurologically and cognitively most severely impaired participants had to be excluded from the analysis.

image

Figure 1.   Cognitive functioning in preschool children with epilepsy. *No psychological assessment data available (n = 3). The level of cognitive functioning was determined to be normal on the basis of medical records, parental reports, and, when available, observations from day care.

Download figure to PowerPoint

Cognitive impairment in relation to epilepsy variables

Some epilepsy-related variables were associated with cognitive functioning (Table 2). Significant correlations were found between cognitive impairment and group, that is, uncomplicated versus complicated epilepsy (r = 0.636, p < 0.001), additional neurologic diagnosis (r = 0.565, p < 0.001), age at onset (r = −0.309, p = 0.01), and MRI (r = 0.258, p < 0.05). However, no significant associations were found between cognitive impairment and seizure frequency, etiology, duration, status epilepticus, most recent EEG, or number of AED.

Table 2.   Cognitive impairment in relation to epilepsy-related variables
 Cognitive impairmentLevel of mental retardation
No (≥70)Yes (≤69)Severe (<34)Moderate (35–49)Mild (50–69)
  1. a inline image = 6.25, p < 0.05.

  2. b inline image = 4.27, p < 0.05.

  3. c Seizure control: Good: Seizure remission >1 year, Partial: Seizures ≤1 per month, Poor: Seizures >1 per month.

Etiologya
 Idiopathic511
 Symptomatic918756
 Cryptogenic1813337
MRIb
 Normal2416538
 Abnormal816556
Seizure type
 Focal1611317
 Generalized1615438
 Unclassified155
Age at onset, months
 1–241221588
 25–4812844
 49–608312
Seizure controlc
 Good1410325
 Partial6422
 Poor1218747
AED
 No AED422
 Monotherapy2219739
 Polytherapy611353

A logistic regression model was used to study the predictors of cognitive impairment. Cognitive impairment was a dependent variable. First, the predictive significance of each independent variable was determined by univariate logistic analysis. Thereafter, the best combination of predictors was computed using a forward stepwise regression model. In the final model age at onset of seizures, etiology, MRI (normal, abnormal), additional neurologic diagnosis (yes, no), and seizure frequency (good, partial, poor) were entered as predictors of cognitive impairment. The overall model predicted about 32% of the cognitive impairment (χ2 = 17.54, d.f. = 5, p = 0.004, Pseudo R2 = 0.32), with age at onset of seizures being the only significant predictor (Wald = 4.52, d.f. = 1, p < 0.05).

Discussion

  1. Top of page
  2. Summary
  3. Purpose
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. References

To the best of our knowledge, there are no other studies available focusing on cognitive impairment in population-based cohorts of preschool children with epilepsy. In most studies (e.g., Rejnö-Habte Selassie et al., 2008), children with additional neurologic impairments and lower IQ (<70) are often excluded. The present population-based study of preschool children with epilepsy included all the 3–6 years 11 months aged children with epilepsy in the hospital district. The prevalence rate 3.2 of this study was consistent with earlier studies conducted in Finland (Eriksson & Koivikko, 1997) and other developed countries (Beilmann et al., 1999; Waaler et al., 2000; Larsson & Eeg-Olofsson, 2006). In the present study, we described the medical and cognitive characteristics of a population-based cohort of preschool children with epilepsy, that is, with no exclusion criteria. The children with uncomplicated epilepsy differed clearly from the complicated group regarding age at onset of epilepsy, duration of epilepsy, and seizure control. It was characteristic of the complicated group that they were younger at onset of seizures, had longer duration, and had higher seizure frequency than the uncomplicated group. Other studies (e.g., Vasconcellos et al., 2001) confirm this to be typical for children with symptomatic or complicated epilepsy.

First, the frequency and the level of cognitive impairment were determined. The mean overall IQ (m = 76.00) of the whole cohort is significantly below normal. In this study, cognitive function was considered normal in 40% and borderline in 9% of the cohort. Moderate to severe mental retardation was found in 28% of the children. These results are consistent with earlier studies (Sillanpää, 1992; Sidenvall et al., 1996; Waaler et al., 2000; Camfield & Camfield, 2007; Berg et al., 2008) suggesting that about 20–57% of children with epilepsy are cognitively impaired. Berg et al. (2008) have recently shown in their large, representative community-based cohort that about 26% of subjects had subnormal cognitive functions and in those with seizure onset before 5 years the percentage was 33%. However, in our study, about 50% had cognitive impairment. The high proportion of participants with low cognitive functioning was somewhat unexpected. As we hypothesized in the light of Mandelbaum and Burack (1997), Bailet and Turk (2000) and Vasconcellos et al. (2001), impaired cognitive functioning was more common among children with complicated epilepsy and within normal range in children with uncomplicated epilepsy. In our study, only one of six children with idiopathic epilepsy had mild cognitive impairment.

Secondly, the associations between the epilepsy-related variables and cognitive functions were studied. The hypothesis was that cognitive impairment in preschool children would be associated with four factors, that is, underlying etiology, age at onset of seizures, seizure duration, and seizure frequency, that other studies have previously shown to contribute to cognitive functions (Battaglia et al., 1999; Vasconcellos et al., 2001; Elger et al., 2004; Cormack et al., 2007). As expected, cognitive impairment was related to age at onset of seizures, abnormal MRI, and additional neurologic diagnosis. A strong relationship between early onset epilepsy and cognitive impairment has been recognized in the literature (Battaglia et al., 1999; Bulteau et al., 2000; Vasconcellos et al., 2001; Chaix et al., 2003; Freitag & Tuxhorn, 2005; Kaaden & Helmstaedter, 2009; Vendrame et al., 2009). Evidently, our study cohort represents children with early onset epilepsy.

Some, for example, Vasconcellos et al. (2001) have demonstrated that early age at seizure onset is a risk factor for cognitive impairment independent of etiology. However, even among this narrow age group there was a significant difference in age at onset of epilepsy between uncomplicated and complicated epilepsy. Complicated epilepsy is closely related to earlier onset of seizures and higher seizure frequency. In addition to early onset, almost half of the children had poor seizure control. Again, seizure control or frequency was poorer in the complicated group. Interestingly, seizure frequency was not associated with level of cognitive impairment, but, this may be partly due to small sample size.

This study provides further evidence of the importance of the age at onset of seizures for cognitive functioning. In addition, in their recent study, Kaaden and Helmstaedter (2009) concluded that cognitive impairment is more related to age at onset of epilepsy than to longer duration or cumulative effects of seizures. In the present study, cognitive impairment was evident in young children (i.e., preschoolers) with relatively short duration of epilepsy supporting the hypothesis that cognitive deterioration or impairment is likely to occur early in the course of epilepsy, as Meinardi and Nunes (1992) and Neyens et al. (1999) have suggested. The developing, immature central nervous system is vulnerable to the effects of damage or harmful internal or external conditions (Aylward, 1997). Therefore, any interruption (e.g., the presence of an active electrophysiologic epileptic focus or repetitive epileptiform generalized discharges) may disrupt the expected developmental course and impede or impair the normal maturation of the central nervous system and cognitive functions (Elger et al., 2004; Hommet et al., 2006; Kaaden & Helmstaedter, 2009). This may be one explanation for the cognitive impairment in young children with epilepsy. Or possibly early seizures affect the brain development and exacerbate subsequent vulnerability of the mature brain to the effects of seizures (Squier et al., 2003; Holmes, 2004). Note also that children with uncomplicated epilepsy may be at risk for neurocognitive difficulties despite normal cognitive capacity (Bailet & Turk, 2000; Henkin et al., 2005; Northcott et al., 2005). According to our earlier studies of this same cohort (Rantanen et al., 2009, 2010), the preschool children with uncomplicated epilepsy demonstrated more problems in neurocognitive functions (e.g., in attention and language) and social competence than their healthy peers.

One of the main benefits of this study is that the focus is on preschool-aged children, a group of children with epilepsy that has been paid little attention in the earlier studies. The strengths of the study also include a targeted and population-based cohort with a narrow age range and the detailed review of the clinical data and epilepsy variables. However, some limitations must be conceded. First, because we focused on a population-based cohort in a narrow age range, the sample size remains small. Due to the methodologic decision to use the population-based cohort, the study group was inevitably heterogeneous, representing the whole range of different types of epilepsy. Because of this, the associations between cognitive impairment and epilepsy-related factors may not be regarded as definite. Failure to find independent effects of MRI abnormalities and etiology in logistic regression may partly reflect inadequate power due to small number of participants. Due to definition used in this study, it is also possible that there may be some overlap between etiology and uncomplicated versus complicated epilepsy. Therefore, the failure of etiology to predict impairment may partly relate to the method of defining etiology. The generalizability of our findings is limited and, therefore, these data must be interpreted with caution.

To conclude, early onset epilepsy is a risk factor for cognitive impairment. Furthermore, the cognitive impairment can already be demonstrated during preschool years in children with relatively short duration of epilepsy. However, due to the general instability of test scores in young children, the lower scores during preschool years may partly be an indication of developmental delay rather than mental retardation. A follow-up study of this cohort would be beneficial to study this and also the association between a longer duration of illness and development of cognitive function. In clinical practice, we should pay more attention to early intervention to improve the subsequent cognitive and psychological outcomes of these children. In addition, further longitudinal, especially prospective studies, are needed to investigate the developmental course of children with epilepsy.

Acknowledgments

  1. Top of page
  2. Summary
  3. Purpose
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. References

This study was supported by Competitive Research Funding of the Pirkanmaa Hospital District and Finnish Epilepsy Research Foundation.

Disclosure

  1. Top of page
  2. Summary
  3. Purpose
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. References

None of the authors has any conflict of interest to disclose. We confirm that we have read the Journal’s position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.

References

  1. Top of page
  2. Summary
  3. Purpose
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. References