Psychiatric symptoms in children prior to epilepsy surgery differ according to suspected seizure focus
Jay A. Salpekar,
Department of Psychiatry and Behavioral Sciences, Children's National Medical Center, George Washington University School of Medicine, Washington, District of Columbia, U.S.A
Center for Neuroscience and Behavioral Medicine, Children's National Medical Center, George Washington University School of Medicine, Washington, District of Columbia, U.S.A
Address correspondence to Jay A. Salpekar, Department of Psychiatry and Behavioral Sciences, Children's National Medical Center, 111 Michigan Ave., NW, Washington, DC 20010, U.S.A. E-mail: email@example.com
Children and adolescents with epilepsy have an overrepresentation of psychiatric illness. However, few studies in pediatrics have characterized specific psychiatric conditions associated with seizure localization. In addition, degree to which psychiatric illness may be more prominent in children refractory to standard medical treatment for epilepsy is not known. The aim of this study was to assess psychiatric symptoms in children with medically refractory epilepsy and ascertain whether symptoms were associated with specific localization.
Case records were reviewed for 40 children with medically refractory epilepsy at the time of their referral for presurgical evaluation. Patients received a clinical psychiatric evaluation and parents completed the Child Behavioral Checklist (CBCL). Seizure localization was verified by pediatric epileptologists, and suitability for surgical procedures was verified by neurosurgical specialists. Groups were compared based on localization of seizure foci, either in the temporal lobe or predominantly extratemporal.
The majority of the sample had psychiatric diagnoses and behavior problems, well beyond the level reported in chronic epilepsy populations. In addition, children with temporal lobe seizure foci had more CBCL behavioral problem categories rated in the clinically significant range, and also were more likely to have clinical diagnoses of depression.
Routine psychiatric evaluation prior to epilepsy surgery may be important for pediatric patients with medically refractory epilepsy. Psychiatric illness, particularly depression, may be especially prominent for those with temporal lobe seizure foci.
Despite the increase in medications available for the treatment of epilepsy, many pediatric patients have seizures that remain refractory to medical treatment, and thus have seizures that persist into adulthood. New surgical techniques offer treatment options for adults and children whose epilepsy is deemed refractory to standard medical approaches (Engel, 1994; Shields, 2004b). Patients who present for presurgical evaluation have been deemed refractory to medical treatment, and on average, have attempted at least three or more courses of drug treatment without satisfactory result (French et al., 2004). Surgical treatment may be regarded as either curative or palliative in terms of reducing the frequency of the most debilitating seizure types (Olson, 2001). Candidacy for a surgical procedure is also determined by epilepsy variables, including discrete localization and surgical accessibility of seizure foci, and proximity to functional tissue such as speech and language regions. However, in addition to these variables, psychiatric evaluation is increasingly recognized as a crucial part of a comprehensive presurgical workup for pediatric patients with epilepsy (Keene & Ventureyra, 2000; Savard & Manchanda, 2000; Danielsson et al., 2009; Souza-Oliveira et al., 2009).
The goals of psychiatric evaluation include diagnostic assessment as well as precise characterization of behavioral symptoms related to preictal or postictal states. Although there are few absolute psychiatric contraindications to epilepsy surgery, severe psychiatric illness may dominate a patient's clinical condition and represent a key component of the overall risk and benefit profile for epilepsy surgery. Recent paradigms of the surgical treatment of epilepsy have expanded outcome measures to include behavioral and developmental function as well as seizure control (Gilliam et al., 2004; Goldstein et al., 2004). One rationale for early consideration of surgical treatment in pediatrics is that timely surgery may enable better recovery of cognitive function and a more typical neuropsychiatric developmental course than is possible with surgical intervention as an adult (Bourgeois et al., 1999; Shields, 2004a). Improvement in cognitive function may improve more with early temporal lobe surgery as compared to extratemporal lobe surgery (D'Argenzio et al., 2011; Skirrow et al., 2011).
Psychiatric problems are known to be overrepresented in adults and children with epilepsy and may be even more prevalent in refractory epilepsy (Inoue & Mihara, 2001). Existing pediatric studies report prevalence in the 20–40% range for psychiatric illness associated with epilepsy (Rutter et al., 1970; Caplan et al., 2005; Jones et al., 2008). The most common psychiatric comorbidities in children include depression, anxiety, and cognitive or learning problems (Brent, 1986; Dunn, 2003; Dunn et al., 2003; Plioplys, 2003; Caplan et al., 2004; Dunn & Austin, 2004); however, psychiatric vulnerability has not been consistently correlated with specific seizure types or localization—key parameters for consideration of surgical treatment. In addition, the psychiatric profile of patients who are refractory to standard medical treatments has not been well characterized in pediatrics (Cankurtaran et al., 2005). A recent study of adult surgical patients found that depression was associated with seizure foci in the temporal lobe—a common focus for surgically treated epilepsy (Sanchez-Gistau et al., 2010). We investigated whether psychiatric problems were more prominent in medically refractory patients with temporal lobe foci versus extratemporal foci. Our hypothesis was that children with temporal lobe seizure foci would have more psychiatric symptoms than those with seizure foci elsewhere in the brain.
At our epilepsy center, patients being considered for epilepsy surgery routinely undergo psychiatric evaluation. Procedures include a full psychiatric diagnostic assessment; medical, neurologic, and psychiatric history; and detailed organic mental status examination. Psychiatric evaluations typically occur on an outpatient basis within a month of the epilepsy team's recommendation for surgery. As a result, children and adolescents with medically refractory epilepsy have broad clinical data regarding psychiatric status.
The sample reflects a consecutive series of pediatric patients with epilepsy who were undergoing routine psychiatric evaluation prior to consideration for epilepsy surgery. All data were obtained prospectively for clinical purposes. Patients were included in the study if epilepsy was the primary diagnosis and psychiatric evaluation was completed. Patients did not receive psychiatric evaluation if communication, age, or low IQ level prevented full participation in the procedures by the patients themselves. All patients were assessed in the context of a presurgical evaluation, given that the epilepsy had been deemed refractory to medication treatment approaches. Recommended surgical procedures included focal resections, extended lobar resections, or hemispherectomy. Although everyone in the sample had been recommended for surgical procedures, a few did not ultimately receive surgical treatment. Reasons for declining surgical treatment included a preference to pursue continued medical options, or resistance to assuming the potential risks of a surgical procedure.
Forty patients were qualified for inclusion in the sample. Demographic details are listed in Table 1. During the period of the review, presurgical psychiatric evaluations were performed in an additional six patients but were excluded because of age (younger than 6 years or older than 17 years), or because additional assessment revealed the full scale IQ to be well below 70. Neurologic history and seizure localization were reviewed and confirmed by board certified pediatric neurologists specializing in epilepsy. Seizure localization was determined by a combination of methods over a 1–2 month presurgical evaluation period. Each patient was assessed initially in terms of ictal clinical characteristics and routine sleep-deprived electroencephalography (EEG) studies. In addition, a prolonged (ictal) video-EEG recording was obtained as well as a high-resolution magnetic resonance imaging (MRI) (1.5 T) using an epilepsy protocol. For most patients, localization was further evaluated during surgery by invasive monitoring, either with cortical surface grids or electrocorticography. Surgical outcome was reviewed, and positive outcomes were viewed as reflecting accurate presurgical determination of seizure foci. Engel outcomes were determined in the sample to the extent that follow-up was available. Ratings of one or two were considered confirmatory for localization as initially suspected. A rating of three was considered confirmatory with focal cortical dysplasia/tumor and known incomplete surgical resection. For the three patients with total or partial hemispherectomies, and for the four patients who did not receive surgical resections, at least 48 h of ictal video-EEG recording or in one case invasive monitoring, was considered for determining localization.
Table 1. Sample demographics
Temporal foci (24)
Extratemporal foci (16)
Average age, years (range)
26 M; 14 F
16 M; 8 F
10 M; 6 F
Average age of onset of epilepsy (years)
Average number of years of epilepsy
Number ultimately receiving surgical procedure
Limited resection (restricted to specific region within a lobe, e.g., amygdalohippocampectomy)
Extended resection (larger scale resection throughout a lobe, e.g., temporal lobectomy)
Initial clinical changes or EEG changes that occurred in the temporal lobe were considered worthy of inclusion in the temporal lobe group. For three patients who had extended resections that included temporal and extratemporal structures, group assignment was based on initial EEG or clinical characteristics. A diagnosis of mesial temporal sclerosis justified inclusion in the temporal lobe group. If the regions resected were small or simply reflected wider margins of resection beyond the region of interest, the initial group determination remained. Group assignments did not change following surgery, but outcomes were recorded. Details of neurologic characteristics in the sample are included in Table 2 (Temporal Lobe Group) and Table 3 (Extratemporal Group).
Patients were categorized as having temporal lobe seizure foci only if the primary seizure focus was strongly suspected to be in the temporal lobe. If the primary seizure focus appeared to be concentrated outside of the temporal lobe, patients were characterized as having extratemporal foci. Based on this standard, 24 patients had temporal lobe seizure foci, and 16 were characterized as having extratemporal lobe seizure foci. Of the 16 patients classified as having extratemporal foci, 11 had foci primarily in the frontal lobe. 4 had foci in the parietal lobe, and 1 had a focus in a posterior (parietal/occipital) region.
Each patient underwent a comprehensive initial psychiatric evaluation and was assigned clinical Diagnostic and Statistical Manual of Mental Disorder (DSM-IV-TR) diagnoses if warranted. For each subject, the same psychiatric consultation team made clinical diagnoses in a consistent manner. A board-certified child and adolescent psychiatrist (JS) specializing in neuropsychiatric evaluations led the evaluation team and confirmed psychiatric diagnoses. Parents completed a Child Behavioral Checklist (CBCL), a 118-item behavior questionnaire, as part of the evaluation. The CBCL (Achenbach, 1991; Achenbach & Rescorla, 2001) is a widely used behavioral questionnaire that generates standardized scores for broadband behavior scales (Total, Externalizing, and Internalizing) and narrowband behavioral scales (Withdrawn, Somatic, Anxiety/Depression, Social, Thought, Attention, Delinquency, and Aggression). Specific narrowband behavioral scales have demonstrated some consistency with Axis I diagnoses such as attention-deficit/hyperactivity disorder (ADHD), depression, and conduct or oppositional defiant disorders. Dimensions of anxiety and depression may be particularly relevant for Axis I diagnoses of depression (Achenbach & Dumenci, 2001). The parent rates the child on problem items using a 0–1–2 (0 = not true, 1 = somewhat or sometimes true, 2 = very true or often true) scale on behaviors in the past 6 months. The range of test–retest value is 0.95–1.00, and the range of internal consistency is 0.78–0.97 (Achenbach & Dumenci, 2001). The CBCL generates T-scores compared to a normative mean accounting for age and gender. For narrowband categories, a T-score of 65+ reflects a problem rating of one and a half standard deviations above the normative mean and is considered clinically “at risk.” A T-score of 70+ reflects a problem rating of two standard deviations above the normative mean and is considered clinically significant. For broadband categories, the T-score thresholds are slightly lower; a T-score of 60+ indicates “at risk” and 64+ indicates clinical significance. Clinicians were not blinded to epilepsy characteristics or psychiatric conditions, but they were blinded to CBCL results at the time of evaluation. Institutional review board approval was obtained for retrospective review of medical records.
The presence of clinical diagnoses and CBCL subcategories with T-scores in the at-risk level (65+ for narrowband and 60+ for broadband), and the clinically significant level (70+ for narrowband and 64+ for broadband) were analyzed. Group comparisons were performed based upon temporal versus extratemporal lobe seizure foci characterization. The number of significant CBCL categories were aggregated among groups and analyzed according to individual subcategory domains as well as for the groups as a whole. Two by two chi-square (χ2) analyses and two-tailed Fisher's exact tests were performed to assess group differences in terms of seizure localization and CBCL narrow-band categories in the clinically significant range. Fisher exact scores enabled 2 × 2 analyses that incorporated the likelihood of odds ratios being either greater or less than one, thereby accounting in part for possible type I and type II errors.
In this sample, 33 (82.5%) of 40 patients had at least one Axis I psychiatric diagnosis based on clinical interview. The most common diagnoses were ADHD (16 [40%] with combined type, inattentive subtype, or Not Otherwise Specified [NOS]), anxiety disorders (13 [32.5%] with Generalized Anxiety Disorder or Anxiety NOS), and depression (8 [20%] with Depressive Disorder NOS). Eleven patients (28%) had two or more diagnoses, including three patients who had previously received psychiatric evaluation and treatment.
At least one CBCL narrowband category in the clinically significant range (T-score of 65+) was present in 31 patients (78%), and 21 patients (53%) had at least two clinically significant CBCL narrowband categories. The most common significant narrowband categories were in domains of social (14 [35%]) and attention (14 [35%]) problems. Also common were somatic (10 [25%]) and thought (10 [25%]) problems. Few patients had significant findings in domains of aggression or rule breaking.
For broadband CBCL categories, 14 (35%) of 40 patients had scores in the clinically significant range (T-score of 60+) for internalizing problems, and 10 (25%) of 40 patients for externalizing problems. The total problem score was clinically significant (T-score of 60+) for 17 (43%) of 40 patients.
Temporal versus extratemporal seizure foci
Both groups had a high frequency of comorbid psychiatric illness. Among the 24 patients with temporal lobe seizure foci, 7 (29%) had anxiety disorders, 8 (33%) had ADHD, and 8 (33%) had depressive disorders. Among the 16 patients with extratemporal foci, 6 (38%) had anxiety disorders, 7 (44%) had ADHD, and only 1 (6%) had a depressive disorder. A clinical diagnosis of depression was common in patients with temporal lobe seizure foci, and trended toward a significant difference (χ2 = 3.937; p = 0.061).
Although clinically significant scores for broadband CBCL categories were common in both groups, temporal foci and extratemporal foci were not significantly different for internalizing problems (χ2 = 3.095; p = 0.101), externalizing problems (χ2 = 2.222; p = 0.263), or total problems (χ2 = 1.381; p = 0.332). However, if patients were selected for having clinically significant scores for either internalizing or externalizing problems, the temporal foci group trended toward significance (χ2 = 4.31; p = 0.054).
Elevated CBCL narrowband category scores were common in both groups, although a greater percentage of the patients with temporal lobe foci had met criteria for clinical significance than did the patients with extratemporal foci (Fig. 1). The largest group differences occurred for two or more categories elevated at T-scores of 65+ (χ2 = 4.708; p = 0.05). Although more patients with temporal lobe foci had two or more categories elevated at T-scores of 70+, this did not reach a level of statistical significance (χ2 = 3.258; p = 0.104).
The two groups were similar in areas of attention problems and social problems. The most common area of concern reported for patients with extratemporal foci was attention problems (Fig. 2). For patients with temporal foci, attention and somatic problems were leading areas of concern (Fig. 3). However, depression-specific CBCL categories (anxious/depressed, withdrawn/depressed) were higher in patients with temporal lobe seizure foci. When both depression categories were analyzed in aggregate, patients with temporal lobe foci had more depressive problems as evidenced by reporting on the CBCL (χ2 = 5.889; p = 0.027).
To account for possible ambiguity of localization determinations, a subanalysis was done for depression categories, eliminating patients with complete hemispherectomies (one temporal and one extratemporal). When these patients were eliminated, the association of temporal lobe foci and depressive problems reported on the CBCL was essentially unchanged (χ2 = 5.982; p = 0.026).
Psychiatric illness is widely present in this sample of children and adolescents with medication-resistant epilepsy. Approximately 80% of the subjects had notable psychiatric symptoms, consistent with rates of psychiatric illness reported in two smaller studies of presurgical children (Szabo et al., 1999; McLellan et al., 2005), and far exceeding the 20–40% rate of psychiatric comorbidity found in chronic pediatric epilepsy (Ott et al., 2003). Although psychiatric diagnoses and symptoms were common regardless of the seizure foci, patients with temporal lobe foci had more psychiatric problems, especially with depression. This finding is intriguing given recent studies implicating the temporal lobe in the pathophysiology of depression. Neuroimaging studies in adults have suggested that mesial temporal lobe structure such as the hippocampus and amygdala have identifiably different size and function in the context of depression (Sheline, 2003). Hippocampal atrophy has been noted in adults with epilepsy and depression, beyond the degree attributable to epilepsy alone (Shamim et al., 2009). Although effects of laterality and specific localization have not been consistently found, depression comorbid with epilepsy has been described as a transactional process where either condition increases the risk for the other (Kanner, 2004). It should also be considered that seizure localization alone may not explain depressive pathophysiology, and some reports suggest that a disordered circuit of neurons possibly extending beyond the temporal lobe may lead to a depressive disorder (Gilliam et al., 2003; Hecimovic et al., 2003). Together with the results of this study, it is plausible to consider that temporal lobe seizure foci may disrupt the healthy function of the amygdala and hippocampus, and may increase the vulnerability for depression in medically refractory pediatric epilepsy.
However, for some psychiatric conditions, group differences were not present between patients with temporal lobe foci versus extratemporal foci. Both groups had high scores on social and somatic problems, which may reflect a global psychosocial or physical impact of medically refractory epilepsy. Disrupted consciousness, physical symptoms, and lifestyle changes may thwart typical social function regardless of seizure foci. Anxiety symptoms also were present in both groups equally, suggesting that anxiety also may be a global phenomenon or at least determined by factors other than specific seizure focus. The etiology of anxiety may be more heterogeneous than that of depressive disorders, with psychosocial and familial factors playing a large role (Wood et al., 2008; Ekinci et al., 2009).
Attention problems and ADHD were prevalent, consistent with existing reports of ADHD as the most common psychiatric comorbidity in children and adolescents with epilepsy (Salpekar & Dunn, 2007). Yet in this sample, attention problems also were not associated with specific seizure foci. The pathophysiology of ADHD is not well established, but recent studies suggest abnormalities in subcortical structures such as the caudate nucleus or cingulate gyrus (Semrud-Clikeman et al., 2006). Another recent study with a large sample suggests that ADHD involves widespread impairments in connectivity between the superior parietal lobe, cerebellum, and orbitofrontal cortex, as well as the ventral striatum and cingulate gyrus (Tomasi & Volkow, 2012). Similar to the findings with anxiety, it may be that the pathophysiology of ADHD is more globally related to epilepsy, determined by wider ranging neural networks, independent of discrete brain regions where seizure foci may be present.
Limitations of this study include sample size, although still larger than in many other reports of presurgical pediatric epilepsy. A larger sample may allow more robust statistical analysis correlating psychiatric symptoms with specific epilepsy variables such as effects of laterality, seizure type, and precise age groups. Semistructured psychiatric diagnostic interview instruments may also help to validate diagnoses. In addition, the majority of the patients had temporal lobe seizure foci, which may be more commonly found in patients with refractory epilepsy who present with pathology that is amenable to surgical treatment. Therefore, behavioral problems that are relevant to temporal lobe localization, for example, affective disorders, may have been overrepresented in the group as a whole. Still, our sample was similar in that all subjects had medically refractory epilepsy and were assessed in the time frame and context of a presurgical evaluation.
Also limiting sample size was the exclusion of patients with IQ <70. This effort was necessary to ensure homogeneity of the sample and to make effective use of standardized rating scales, the utility of which is less certain in patients with intellectual disability (Masi et al., 2002). Significant psychopathology may be associated with patients with intellectual disability and epilepsy; however, assessment approaches differ in terms of symptom characterization (Arshad et al., 2010). Despite these limitations, we view this investigation as noteworthy in that patients with temporal lobe foci appear to have more severe behavioral issues, especially in terms of depressive disorders.
Identification of precise seizure foci may also present a limitation to the study. Use of video-EEG studies, invasive monitoring, and surgical outcomes provided confidence in suspected localizations. Group determinations were based on primary seizure foci, but given the heterogeneity of the epilepsy, epileptogenic zones may not be clearly identifiable, even with high quality evaluative tools. Outcomes were largely positive, but some limited outcomes were attributable to incomplete resections—proximity to eloquent tissue, focal cortical dysplasia that may have been undetected, incomplete lateralization, or multiple seizure foci. In addition, some resections included both temporal and extratemporal structures, notably in two patients with complete hemispherectomies. Therefore, surgical outcome in these cases may be less clearly attributed to specific pathology in temporal versus extratemporal brain regions. In this sample, the subanalysis eliminating these patients did not change the finding of temporal lobe foci being more associated with clinically significant depression scores on the CBCL. Despite these limitations, the constellation of MRI and surgical outcome joined to EEG provides reasonable assurance of localization in this sample, sufficient to posit temporal lobe pathology as more associated with psychiatric symptoms.
Given that the majority of children and adolescents in the sample had behavioral issues, routine psychiatric evaluation may be essential in the comprehensive care of presurgical pediatric patients with epilepsy. Regardless of seizure localization, psychiatric evaluation may be indicated for most patients with medically refractory epilepsy. In most cases, the psychiatric issues eluded clinical referral, and the only contact with a child psychiatrist was in the context of the presurgical evaluation. If temporal lobe pathology is more complex in terms of psychiatric comorbidity and disrupted quality of life, then epilepsy centers need to be cognizant of increased mental health treatment needs in order to properly care for these patients. Further study that focuses on the outcome of surgery on psychiatric issues in addition to epilepsy variables will be key areas of future investigation.
No grants or financial aid were received to support this study.
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.