Cognitive changes following surgery in intractable hemispheric and sub-hemispheric pediatric epilepsy
- First Online:
- Cite this article as:
- Thomas, S.G., Daniel, R.T., Chacko, A.G. et al. Childs Nerv Syst (2010) 26: 1067. doi:10.1007/s00381-010-1102-5
- 237 Views
The objectives were to study the short and longitudinal changes in the cognitive skills of children with intractable epilepsy after hemispheric/sub-hemispheric epilepsy surgery.
Sixteen patients underwent surgery from September 2005 until March 2009. They underwent detailed presurgical evaluation of their cognitive skills and were repeated annually for 3 years.
Their mean age was 6.6 years. Epilepsy was due to Rasmussen’s encephalitis (n = 9), Infantile hemiplegia seizure syndrome (n = 2), hemimegalencephaly (n = 2), and Sturge Weber syndrome (n = 3). Fourteen (87.5%) patients underwent peri-insular hemispherotomy and two (12.5%) underwent peri-insular posterior quadrantectomy. The mental and social age, gross motor, fine motor, adaptive, and personal social skills showed a steady increase after surgery (p < 0.05). Language showed positive gains irrespective of the side and etiology of the lesion (p = 0.003). However, intelligence quotient (IQ) remained static on follow-up. Patients with acquired pathology gained more in their mental age, language, and conceptual thinking. Age of seizure onset and duration of seizures prior to surgery were predictive variables of postoperative cognitive skills.
There are short- and long-term gains in the cognitive skills of children with intractable epilepsy after hemispherotomy and posterior quadrantectomy that was better in those patients with acquired diseases. Age of seizure onset and duration of seizures prior to surgery were independent variables that predicted the postoperative outcome.
KeywordsIntractable epilepsyEpilepsy surgeryCognitive declineCognitive outcomeChildren
“Intractable epilepsy” in adults was previously defined as a failure to respond to at least two antiepileptic drugs (AEDs) given over at least a 2-year period . A single definition for “intractable” epilepsy cannot suit all situations as definitions of intractability are individualized to the patient. Of these patients deemed to be intractable, approximately 50% are estimated to have surgically remediable epilepsy [2, 3]. “Hemispheric epilepsy” refers to epileptiform activity in all four lobes of one hemisphere, and when it involves more than two lobes of the brain, it is termed “sub-hemispheric epilepsy” [2, 4].
Children with refractory epilepsy are at considerable risk for cognitive impairment [5–10] as well as school failure , behavioral and mental health problems [6, 12–14], and overall compromised quality of life . The numerous risk factors for epilepsy-related cognitive decline discussed in literature are the mere presence of seizures , seizure-related variables such as age of onset, frequency and/or severity, duration , underlying pathology , AEDs [6, 10, 11, 14], and various psychosocial factors  often called the “burden of epilepsy” .
Neuropsychological assessment in both children and adults is important before and after surgery. Yet, there is limited neuropsychological outcome data available in children with hemispheric/sub-hemispheric epilepsy. The first goal of pre- and postsurgical neuropsychological assessments in children is the evaluation of cognitive development (development delay, stagnation, or deterioration) and, second, to look for the possible effects of both brain injury, epileptic disorder, and chronic treatment with AEDs. Moreover, it allows the identification of the functional consequences of surgical procedure and, possibly, to single out some predictive factors of cognitive development. This study was designed and conducted to add information to the existing literature on the cognitive outcome of these children following surgery and the variables that could predict the outcome.
Patients and methods
Children with IE were evaluated at the preoperative epilepsy meeting. The evaluation included a study of the seizure semiology, neurological examination, multiple electroencephalogram examinations (video telemetry), MRI, and neuropsychological evaluation. The focus of the examination was to identify a surgically remediable epilepsy syndrome with good electro-clinico-radiological concordance. A total of 16 patients underwent disconnective epilepsy surgery for hemispheric/sub-hemispheric epilepsy from September 2005 until March 2009. All surgeries were done by one surgeon (RTD). Their ages ranged from 8 to 159 months (mean age = 79.13 months; 6.6 years). There were 10 (62.5%) males and six (37.5%) females. The mean age of seizure onset was 3.4 years (range, neonates–11.5 years). The mean duration of epilepsy was 3.2 years (range, 3 months–10.5 years). Eleven of them (68.7%) were on more than two AEDs. The etiologies of epilepsy were Rasmussen’s encephalitis (n = 9, 56.25%), Infantile hemiplegia seizure syndrome (n = 2, 12.5%), hemimegalencephaly (n = 2, 12.5%), and Sturge Weber syndrome (n = 3, 18.75%). Fourteen (87.5%) patients underwent peri-insular hemispherotomy for lesions causing hemispheric epilepsy and two (12.5%) underwent peri-insular posterior quadrantectomy for lesions causing sub-hemispheric epilepsy. Three (18.75%) patients were followed up for 3 years, four (25%) patients for 2 years, and nine (56.25%) patients for 1 year.
A prospective, longitudinal research design was used to document the changes in the cognitive ability of these children. The assessments were done during the presurgical period evaluation and annual follow-up visits. An experienced rehabilitation psychologist administered the Gesell’s Developmental Schedule or Binet–Kamat Scale of Intelligence to the child and Vineland Social Maturity Scale (VSMS) to the caregivers [16–18]. All cognitive assessments were carried out by a single rehabilitation psychologist to avoid inter-rater reliability-related issues.
As part of the data analysis, preliminary checks of skewness verified that our data were suitable for nonparametric analysis. The analyses were carried out at three levels. Firstly, the descriptive statistical analyses included mean and standard deviations for describing the participants’ characteristics. Secondly, Chi-square test with Yates correction, Wilcoxan Signed Ranked test, and Mann–Whitney U tests were used to compare the pre- and postsurgical cognitive variables. Finally, based on the results of the first two levels of analyses and review of literature, some variables related to cognitive gain were attested in a regression model to determine the predictors of cognitive gain. Multiple linear regression yielded regression coefficient for continuous variables, whereas, multiple logistic regression yielded odds ratio and 95% confidence interval for the dichotomized main outcome variable of mental age. A significance level of 0.05 and two-tailed tests were used unless otherwise noted because of the nature of the study hypotheses. Data were analyzed using the software package of SPSS (version 16).
The average number of AED’s used before surgery was 3.0, which reduced to 1.1 after surgery. The mean loss of schooling due to seizures was 1.2 years. Fifteen children (93.75%) had Engel’s Class I outcome or complete seizure freedom and one child had Class II (>90% seizure reduction) outcome. Analysis of data of the published series of hemispherotomy  and our present series showed similar seizure outcomes.
Cognitive gain during follow-up visits
Overall cognitive gain during follow-up visits (N = 16)
Mental age (months)
Non mean memory
Cognitive outcome by etiology
Cognitive status during the last follow-up visit (3rd) by etiological groups
Infantile hemiplegia seizure syndrome
Sturge Weber disease
Statistics df, F
Non mean memory
28.5 (19.0) (0.0)
Cognitive status during the last follow-up visit (3rd) by etiological groups (acquired vs congenital) excluding hemimegalencephaly
Acquired (Rasmussen’s encephlitis) N = 9
Congenital (infantile hemiplegia + Sturge Weber) N = 5
Non mean memory
Cognitive outcome by side of surgery
The results of cognitive changes of the entire sample at the third follow-up with respect to side of surgery showed no significant differences in cognitive domains. Further analysis of the group with Rasmussen’s encephalitis (n = 9, 56.25%) alone also did not show any difference in cognitive outcome with respect to side of surgery.
Predictors of gain in mental and social age of children at 1-year follow-up
Predictors of gain in mental and social age of children (N = 16) at one-year follow-up
Side of surgery
Age of onset of seizures
Duration of seizure
Preoperative VSMS scores and postoperative improvement in cognition
Changes in vineland social maturity scale scores (adaptive skills) in relation with their preoperative scores (N = 16)
% Impairment in preop VSMS scores
N = 16
No. of patients who had gains in VSMS over 3 years (N = 7)
No. of patients who were static in VSMS over 3 years (N = 9)
No. of patients who had losses in VSMS over 3 years (N = 0)
Gd 2 (25–50%)
Gd 3 (50–75%)
Gd 4 (>75%)
There is paucity of literature on the cognitive outcome in children undergoing surgery for intractable hemispheric epilepsy. We designed and conducted this study to document the short- and long-term cognitive outcome in this group of patients. Our results showed statistically significant gains in the domains of language and adaptive behavior in the mean follow-up of 22 months after surgery with a 93.75 % Engel’s Class I outcome.
When we compared cognitive test scores on the basis of etiology, we found a statistically significant improvement in the domains of language with RE scoring the highest followed by SW, IHSS, and HM. These findings were similar to Pulsifer et al.  who also demonstrated that patients with RE and IHSS score higher in general intelligence and language functions than cortical dysplasia. But there was no appreciable difference between the RE and IHSS groups in their study . Furthermore, we also found that patients with acquired disorders (RE) performed better postoperatively than the congenital group (IHSS + SW) in domains of language and conceptual thinking. Thus, hemispherotomy appears to halt the ongoing cognitive decline seen in patients with RE. Because the natural history of RE is that it is a progressive disease, the underlying basis for this phenomenon is not known, though others have reported similar findings [21, 22]. Interestingly, we found good language gains irrespective of side, and no deterioration in language functions were noted following left-sided hemispherotomies when compared to the right side. Even within the RE group, there was no difference in language performance between the right and left hemispherotomy groups. Notwithstanding the difficulty in performing the Wada test in children to locate the language dominant hemisphere, our findings indicate that language function in children with dominant hemispheric epilepsy seems be taken over by the nondominant hemisphere. These findings are corroborated by others [23–25] but, however, are in contrast to the findings of Pulsifer et al.  whose right hemispherectomy RE patients scored significantly higher in receptive and expressive language than their counterparts who underwent left hemispherectomy. It must be noted that in their patients, these language differences by side were observed preoperatively as well as at follow-up and so are most likely due to the original disease pathology. Studies on reorganization after a hemispherotomy showed that ipsilateral representations were topographically different from contralateral representations, with ipsilateral representations occupying more anterior and lateral locations than contralateral representations in the healthy hemisphere .
Duchowny et al.  suggested that in children with congenital disorders causing IE, there is limited scope for language to relocate to the opposite hemisphere. This could be due to the widespread bilateral cortical damage in these patients in utero as opposed to those children who acquire a disease after a normal initial brain development. The period of rapid language development is between 2 and 5 years of age, with slower development continuing through puberty. This explanation is supported by others  who state that when brain damage occurs during this “critical” period, language functions could shift from one hemisphere to the other, with minimal consequences in terms of functional language skills. Preservation of language functions may be associated with the recruitment of intact regions within the damaged left hemisphere or of homotopic regions in the right hemisphere, or of both . There is also a suggestion that language reorganization is more frequent in stroke and inflammation as compared to developmental process and tumors .
Social and adaptive outcomes
Smith and Elliot  report little discernable change in cognitive function 1 year after pediatric epilepsy surgery while Pulsifier et al. , in a series of 71 patients, found that their adaptive skills remained constant over a period of 5.4 years during follow-up. This is surprising considering that 1 year should be long enough for reconfiguration of the individual family functioning and for restitution at the level of brain plasticity to occur . Our patients showed definite gains in VSMS scores in subsequent follow-up after surgery (Table 1). The lack of significant improvement in other studies may be related to their patients having higher preoperative adaptive scores that did not change much after surgery. In general, children with IE in our country are not exposed adequately to environmental stimuli as schooling is discontinued in the presence of seizures. With good postoperative seizure control, the child’s attention capacities increase, and they benefit from environmental input and engage in social interaction with their family and peers. An improved school attendance improves their adaptive skills that in turn helps the child deal effectively with environmental stimuli and substantially improves their performance in activities of daily living. Enhanced social interaction with a concomitant decrease in internalization of symptoms during follow-up after surgery has been shown to have a positive effect on quality of life . Several studies apart from ours indicate that quality of life measures paralleled the improvement in seizure control . In this study, we also found that patients with mild or moderate impairment in social maturity had a significant improvement in their VSMS scores after surgery. Although patients with severe social maturity impairment had little or no improvement in their social skills after surgery, it is noteworthy that all had a halt in cognitive decline further justifying the rationale for surgery in these patients in whom the natural history of their illnesses dictate a progressive deterioration.
Changes in DQ/IQ
Our results showed that though DQ/IQ scores only showed a positive trend, the four domains of DQ assessment (gross motor, fine motor, adaptive, and personal social skills) showed definite statistically significant gains in scores. DQ is performed in children less than 6 years of age. Hence, a positive trend in its subgroup domains indicates that surgery before 6 years of age may be more effective than after the age of six. However, more data and studies are required to confirm this hypothesis. The majority of patients in most reported series demonstrated no apparent change in DQ/IQ after a median follow-up of 2 to 3 years [20, 23, 31, 32]. Our results showed a steady gain in mental age after surgery which was statistically significant. The mental age gain would be expected to be a part of normal chronological development and, hence, cannot be commented upon as part of a cognitive domain.
Predictors of postoperative cognitive outcome
Neuronal disruption sustained earlier in life has a more impairing effect on problem solving and psychometric abilities than does brain damage that occurs after a longer period of normal growth and development . Onset of intractable epilepsy within the first 24 months of life is a significant risk factor for mental retardation, especially if seizures occur daily [14, 33]. Most of these children have a low DQ/IQ to begin with, and the improvement in cognitive skills after surgery is poorer in these children. However, Thomson and Duncan found no significant relation between cognitive change and age at onset and, thus, did not support the hypothesis that early age at onset is a risk factor for decline . Our study showed that age of seizure onset was a predictive variable for pre- and postoperative cognitive status. Older age of onset of seizures showed positive mental and social age gains at follow-up. This finding is also supported by other studies [1, 15]. Furthermore, a shorter duration of seizures prior to surgery is predictive of positive mental and social age gains in our group of patients, a finding that is in keeping with those of Freitag and Tuxhorn  and Steinbok et al .
In our experience, over 90% of children with sub-hemispheric and hemispheric epilepsy syndromes achieve an excellent seizure outcome following epilepsy surgery. We found a significant gain in the language, meaningful memory, and social age in the immediate postsurgical period which continued into the second and third years of follow-up. In addition, patients with Rasmussen’s encephalitis perform better on tests of cognition as compared to congenital diseases particularly with regard to language function irrespective of the side of hemispherotomy. Following seizure freedom, improvement of function in the residual brain occurs that in turn leads to a “catch up” of adaptive and social functions and quality of life.
We would like to acknowledge Prof. Jean Guy Villemure who mentored the surgical epilepsy program in Christian Medical College, Vellore, India. He was a constant source of support and encouragement throughout the establishment of this program. We also thank him for reviewing the manuscript of this research project.