FULL-LENGTH ORIGINAL RESEARCH
A population-based study of long-term outcome of epilepsy in childhood with a focal or hemispheric lesion on neuroimaging
Address correspondence to Elaine Wirrell, MD, Department of Pediatric Neurology, Mayo Building, 16th floor, Mayo Clinic, 200 First Street SW, Rochester, MN 55905, U.S.A. E-mail: firstname.lastname@example.org
Purpose: To evaluate long-term seizure outcome in children with epilepsy and a focal or hemispheric neuroimaging abnormality.
Methods: All children (<18 years and residing in Olmsted County, Minnesota) with new-onset epilepsy diagnosed between 1980 and 2004 and a single focal lesion on neuroimaging were identified by review of the Rochester Epidemiologic Project database. Outcomes were divided into three categories: (1) seizure freedom for 1 or more years at last follow-up, (2) ongoing seizures but not medically intractable epilepsy, and (3) medically intractable epilepsy or undergoing epilepsy surgery. We also evaluated the proportion who achieved seizure control without surgical intervention and whether lesion type predicted intractability.
Key Findings: Of the 359 children with newly diagnosed epilepsy, 37 (10%) had a focal or hemispheric lesion on neuroimaging. Median age of diagnosis was 89 months (25th percentile 26 months, 75th percentile 142 months) and at follow-up was 137 months (25th percentile 95 months, 75th percentile 211 months). Eighty-three percent of children with malformations of cortical development, 67% with mesial temporal sclerosis, 33% with encephalomalacia, and 50% with vascular malformations had intractable epilepsy at follow-up or underwent resective surgery for medically intractable epilepsy. Among the different etiologies, presence of encephalomalacia predicted the lowest likelihood of medical intractability or undergoing surgery (p < 0.01). At final follow-up, 24 (65%) of our entire cohort was seizure free. Following surgery, seizure freedom was achieved in 80% with mesial temporal sclerosis, 67% with encephalomalacia, 67% with vascular malformation, and 50% with malformations of cortical development. There was no statistically significant difference between the different etiologies on neuroimaging and seizure freedom after surgery. Twelve children (32%) achieved seizure freedom with medical management alone.
Significance: Focal lesions on neuroimaging confer a high risk of medical intractability among children with new-onset epilepsy. However, 32% of this cohort achieved seizure remission with medical management alone, including 58% with encephalomalacia and 33% with mesial temporal sclerosis.
Epilepsy is one of the most frequent neurologic disorders affecting the pediatric population. Advances in neuroimaging have made a major impact in the evaluation and management of patients with focal-onset seizures, with high-resolution magnetic resonance imaging (MRI) showing a structural abnormality in a significant number of patients previously labeled as “cryptogenic” (Von Oertzen et al., 2002; Goyal et al., 2004). The most common potentially resectable lesions detected on neuroimaging in children with epilepsy include malformations of cortical development, focal encephalomalacia or gliosis, mesial temporal lobe sclerosis, tumors, and vascular malformations (Woermann & Vollmar, 2009).
Although many children achieve seizure control with medication, approximately 10–40% have medically intractable epilepsy (Camfield et al., 1993; Hauser et al., 1998; Kwan & Brodie, 2000; Berg et al., 2001; Dlugos et al., 2001). Such patients may benefit from surgical resection, with seizure freedom in 60–80% with unilateral mesial temporal lobe epilepsy or tumors and in 40–70% with malformations of cortical development or dual pathology (Sisodiya, 2000; Dlugos et al., 2001; Wieser et al., 2003; Nakase et al., 2007; Spencer & Huh, 2008; Roper, 2009).
Surgical series and hospital-based studies have reported that focal lesions on neuroimaging correlate with high rates of intractability. However, such studies are biased toward selecting patients who have poorly controlled seizures (Semah & Ryvlin, 2005). No study has assessed long-term outcome in a population-based cohort of children documented to have a focal lesion on neuroimaging at initial diagnosis of epilepsy. One prospective study in patients seen at a single epilepsy center examined the success of medical therapy alone in patients with localization-related epilepsy and found that the nature of lesion predicted control. The poorest control was seen with mesial temporal sclerosis (42%) followed by malformations of cortical development (54%), tumor (63%), infarct (67%), and vascular malformation (78%) (Stephen et al., 2001). In another hospital-based observational study of adults, Semah et al. (1998) also found mesial temporal sclerosis to be a poor prognostic factor.
The aim of this population-based, retrospective study was to evaluate what proportion of children with epilepsy and a potentially resectable lesion on neuroimaging (focal or hemispheric) achieved seizure control without surgical intervention and whether lesion type predicted intractability.
Cases were ascertained by screening of the complete diagnostic indexes of the Rochester Epidemiology Project. These indexes include inpatient diagnoses, as well as diagnoses at the time of outpatient and emergency room visits at all medical care facilities in Olmsted County, MN. All charts were screened using a diagnostic rubric, which included all seizure and convulsion diagnosis codes, and all identified charts were reviewed by a pediatric epileptologist. We identified all children aged 1 month through 17 years with new-onset epilepsy diagnosed while residing in Olmsted County, Minnesota, between 1980 and 2004.
Epilepsy was defined as a predisposition to unprovoked seizures (Fisher et al., 2005). Most subjects had two or more unprovoked seizures. However, patients with a single unprovoked seizure who had evidence of an enduring alteration of the brain that increases the likelihood of further seizures (Fisher et al., 2005), and who were commenced on antiepileptic drug treatment, were also included. An abnormal neurodevelopmental examination, focal abnormality on brain imaging, initial presentation in status epilepticus, or specific EEG findings (epileptiform discharge, intermittent rhythmic focal delta activity) were considered indicative of an enduring alteration of the brain that increases the likelihood of further seizures. Patients who were treated after a single seizure, but who lacked any of the preceding features were excluded. We included children who had two afebrile seizures occurring within 24 h, as these children likely have epilepsy (Camfield & Camfield, 2000). Children presenting with acute symptomatic seizures alone (defined as seizures in close temporal association with an acute neurologic insult) were excluded. Similarly, children who had only febrile seizures were excluded. Children with neonatal seizures were included only if their seizures recurred >28 days after birth.
Classification of neuroimaging studies and patient selection
Lesions detected on neuroimaging were categorized by location (generalized, multifocal, hemispheric, frontal, temporal, parietal, and occipital) and presumed etiology (malformation of cortical development, mesial temporal lobe sclerosis, tumors, vascular malformation, infarct, or other). Neuroimaging reports for all cases were available, and where possible, original neuroimaging studies were reviewed by two epileptologists (EW and GC) for confirmation of lesion location and etiology. Only patients with single focal or hemispheric lesion on neuroimaging [brain computed tomography (CT) or MRI] believed to be responsible for their epilepsy were selected for study. Mesial temporal lobe sclerosis was defined on the MRI as hippocampal formation atrophy and an increased mesial temporal T2-signal intensity. Hippocampal formation atrophy is most obvious using the T1-weighted gradient-echo image in the oblique-coronal plane. The high signal intensity alteration can be identified using T2-weighted imaging or the fluid attenuated inversion recovery (FLAIR) sequence in the oblique-coronal plane (Cascino, 2008). Malformation of cortical development was defined on the MRI as focal cortical thickening, blurring of the gray–white matter junction, and hyperintensity (on T2-weighted images) of subcortical white matter tapering toward the ventricle (Colombo et al., 2003).
Chart review and data analysis
Medical records of selected cases were reviewed to determine outcome with or without epilepsy surgery at last follow-up. Outcomes in each of the etiologic categories were stratified into three groups: (1) seizure freedom for ≥1 year at final follow-up (on or off antiepileptic medications), (2) ongoing seizures but not medically intractable, and (3) medically intractable epilepsy or undergoing epilepsy surgery. Medically intractable epilepsy was defined as having seizures more frequently than every 6 months during the year prior to final follow-up, and failing two or more antiepileptic drugs for lack of efficacy.
All data entry and statistical analysis was performed using PASW Statistics 18.0 (SPSS Inc., Chicago, IL, U.S.A.). Correlations were examined using the Fisher’s exact test (two sided table). p-Values <0.05 were considered statistically significant.
This study was approved by the institutional review board of Mayo Clinic, Rochester.
Three hundred fifty-nine children were newly diagnosed with epilepsy between 1980 and 2004 while residing in Olmsted County. The mode of onset and etiologies of seizures in this population has been previously reported (Wirrell et al., 2011). All children except three were seen on at least one occasion by a child neurologist who confirmed the diagnosis of epilepsy. Of the remaining three, two were diagnosed with epilepsy by a general neurologist and one by a pediatrician. Prophylactic antiepileptic drug therapy was commenced in 339 patients, whereas 20 were never treated. Forty-nine were started on medication after their first afebrile seizure, and of these, 34 had subsequent seizures despite medication. Fifteen had only a single seizure, but had one or more features predicting higher rate of recurrence (specified in Methods) and were commenced on antiepileptic drug therapy without recurrence.
MRI became available in our center in 1984. Prior to that, neuroimaging was performed with CT scanning, which arguably could have missed important epileptogenic lesions such as mesial temporal sclerosis or certain malformations of cortical development. Over the period of our study, use of MRI in our entire cohort of new-onset pediatric epilepsy increased, being done in 17 (38%) of 45 patients diagnosed from 1980–1984, 33 of (53%) 62 diagnosed from 1985–1989, 47 (67%) of 70 diagnosed from 1990–1994, 73 (82%) of 89 diagnosed from 1995–1999, and 77 (83%) of 93 diagnosed from 2000–2004 (p < 0.01). Overall, in our cohort of 359 cases of new-onset pediatric epilepsy, neuroimaging was performed in 329 (92%) cases, 247 of whom had MRI, and 82 of whom had CT only. Of those who had MRI, 40 had generalized or multifocal abnormalities and 36 had either focal or hemispheric abnormalities. Of those who had CT scans only, 14 had generalized or multifocal abnormalities and only 1 had a focal or hemispheric abnormality. In summary, 37 (10%) were found to have a focal or hemispheric lesion on neuroimaging responsible for their epilepsy.
Our cohort included 25 male and 12 female patients. Median age of diagnosis was 89 months (25th percentile 26 months, 75th percentile 142 months) and at follow-up was 137 months (25th percentile 95 months, 75th percentile 211 months). The clinical details of the patients at presentation and last follow-up are listed in Table 1.
Table 1. Detailed clinical characteristics of the 37 patients at onset and last follow-up
| Examination at presentation |
|20 – Normal|
|17 – Abnormal|
| 13 – Hemiparesis|
| 2 – Quadriparesis|
| 2 – Clumsiness, hypotonia, extrapyramidal|
| Seizure type at presentation |
|11 – Secondary generalized tonic–clonic seizures|
|31 – Partial seizures|
| 9 – Status epilepticus at presentation|
| Seizure frequency at last follow-up |
|24 – Seizure-free|
| 4 – Seizures every 6–12 months|
| 2 – Seizures every 3–6 months|
| 7 – Seizures more than every 3 months|
| Number of antiepileptics at last follow-up |
|12 – Off all medications|
|16 – On one antiepileptic|
| 8 – On two antiepileptics|
| 1 – On three antiepileptics|
| Seizure type during last year of follow-up in patients that were not seizure-free (n = 13) |
| 6 – Secondary generalized tonic–clonic seizures|
|11 – Partial seizures|
Original imaging studies were reviewed in all but four cases, as they were no longer available. Etiologies on imaging included encephalomalacia (n = 12), malformations of cortical development (n = 6), mesial temporal sclerosis (n = 9), vascular malformations (n = 6) [cavernous hemangioma (n = 2), arteriovenous malformation (n = 2), thrombosed arterial aneurysm (n = 1), Sturge-Weber syndrome (n = 1)], tumor (n = 2) [ganglioglioma (n = 1), fibrillary astrocytoma (n = 1)], and other (n = 2) [hypothalamic hamartoma (n = 1), calcified granuloma (n = 1)]. Table 2 summarizes the epilepsy course of patients with each presumed etiology.
Table 2. Detailed description of epilepsy course in different etiologies
|Median follow-up (months)||131||234||150||108||129|
|25th percentile, 75th percentile||(86, 177)||(133, 260)||(85, 189)|| ||(69, 176)|
|Median age at diagnosis (months)||35||115||85||51||134|
|25th percentile, 75th percentile||(<1, 150)||(67, 173)||(31, 121)|| ||(5, 191)|
| Surgical treatment ||4/6 (67)||5/9 (56)||3/12 (25)||2/2 (100)||3/6 (50)|
| Seizure-free||2/4 (50)||4/5 (0)||2/3 (67)||1/2 (50)||2/3 (67)|
| Intractable||2/4 (50)||0||1/3 (33%)||1/2 (50)||1/3 (33%)|
| Median follow-up (months)||131||260||191|| ||122|
| Medical treatment alone ||2/6 (29)||4/9 (44)||9/12 (75)||–||3/6 (50)|
| Seizure-free||1/2 (50)||3/4 (75)||7/9 (77)|| ||1/3 (33)|
| Intractable||1/2 (50)||1/4 (25)||1/9|| ||0|
| Median follow-up (months)||165||135||149|| ||135|
Of the six patients with malformations of cortical development, 5 (83%) were either medically intractable or underwent surgery. Of the four treated surgically, at final follow-up, two were seizure-free and two had ongoing seizures. Of the two treated with medical therapy alone, one was seizure-free and the other had medically intractable epilepsy.
Six (67%) of nine patients with mesial temporal sclerosis were either medically intractable or underwent surgery. At final follow-up, of the five treated surgically, four were seizure-free. Of the four treated with medical therapy alone, three were seizure-free, and the other had medically intractable epilepsy.
Only 3 of the 12 patients with encephalomalacia underwent surgery and of these two were seizure-free at final follow-up. Of those medically managed, 7 of 9 ultimately achieved seizure freedom, one had ongoing, but not intractable seizures, and one had medically intractable epilepsy.
Both patients with tumor had surgery for tumor removal. At last follow-up, one achieved seizure freedom off antiepileptic medications, whereas the other remained intractable.
Of the six patients with a vascular malformation, three had surgery. Of those treated medically, one was seizure free and neither of the remaining two had developed medically intractable epilepsy.
Following surgery, seizure freedom at final follow-up was achieved in 80% with mesial temporal sclerosis, 67% with encephalomalacia, 67% with vascular malformation, and 50% with malformations of cortical development.
Twenty-one of 37 subjects (57%) were either medically intractable at final follow-up or had undergone epilepsy surgery. Compared to the other etiologies, encephalomalacia was significantly less likely to correlate with this outcome (p < 0.012).
In our population-based cohort of new-onset pediatric epilepsy, 10% had a focal neuroimaging abnormality that was felt to be responsible for their epilepsy. Encephalomalacia was the most common abnormality (32%), followed by mesial temporal sclerosis (24%), malformation of cortical development (16%), vascular malformations (16%), and tumor (5%). In previous surgical series of children with epilepsy, malformations of cortical development have been found to be the most common cause of medically intractable focal onset epilepsy (Lerner et al., 2009). The difference likely relates to our methodology. We identified a population-based cohort of newly diagnosed children with epilepsy, then selected for study both medically and surgically treated patients with lesional neuroimaging. Of this cohort, we found that the presence of encephalomalacia has the most favorable outcome and predicts the least likelihood of being intractable or undergoing surgery.
Previous studies have reported that mesial temporal sclerosis is associated with a high likelihood of intractable epilepsy (Semah & Ryvlin, 2005). Surprisingly, in our cohort, one third of patients with this etiology were seizure-free at final follow-up with medical treatment alone after prolonged follow-up. Similar to prior reports, we found that surgical outcome is favorable with this etiology, with 80% achieving seizure-freedom (Shaefi & Harkness, 2003; Wieser et al., 2003; Mihara et al., 2004; Nakase et al., 2007; Smyth et al., 2007; Spencer & Huh, 2008).
Surgical outcomes are typically less favorable for malformations of cortical development, with reported rates of seizure freedom of 49%, 44%, and 33%, respectively, at 1, 2, and 5 years in one retrospective study (Phi et al., 2010). These rates are comparable to the 50% rate of seizure freedom in our cohort of children with malformations of cortical development who underwent surgery.
Not surprisingly, the majority of children (57%) in our cohort with a focal imaging lesion either underwent surgery (n = 18) or had medically intractable epilepsy at final follow-up (n = 3). This proportion is significantly higher than seen in our overall cohort of children with new-onset epilepsy, where only 20% had either undergone epilepsy surgery or had medically intractable epilepsy at final follow-up (Wirrell et al., 2011). What is notable, however, was that poor outcome was not experienced by all children with focal lesions. Nearly one-third achieved seizure freedom with medical management alone, including 33% with mesial temporal sclerosis and 58% with encephalomalacia. Although these finding were not statistically significant among the different etiologies, the trends are striking and deserve further investigation utilizing larger population-based studies.
Over the course of our study period from 1980–2004, significant improvements in neuroimaging were made, with the advent of MRI, which became available in our center in 1984. Prior to that, neuroimaging was performed with CT scanning, which arguably could have missed important epileptogenic lesions such as mesial temporal sclerosis or certain malformations of cortical development. However, as most of our cohort was followed longitudinally, those whose seizures were poorly controlled ultimately underwent MRI.
Our study is unique given that it is a population-based study of children with epilepsy with focal or hemispheric lesions on neuroimaging with long-term follow-up. Unlike previous studies on this population that have been either surgical series or hospital-based studies, our study cohort is free from selection bias. The major weakness of our study is small sample size. Such size limits our ability to detect statistically significant difference in the outcome of children among the different etiologies.
We conclude that although finding a focal lesion on neuroimaging in a child with newly diagnosed epilepsy portends a higher risk of medical intractability, outcome is not always bleak. Medical management alone was associated with seizure freedom in nearly one third of cases, particularly if the etiology is encephalomalacia. Larger population-based studies are needed to confirm whether lesion etiology predicts outcome.
Dr. E. Wirrell serves on the Editorial Boards for Epilepsia, Canadian Journal of Neurological Sciences, and Journal of Child Neurology. Dr. G. Cascino—Research Grant: Neuro Pace, Inc. (coinvestigator); Honoraria: American Academy of Neurology; Research Grant: National Institutes of Health R01NS053998 (principal investigator); Associate Editor: Neurology (published by the American Academy of Neurology). The remaining authors have no relevant relationships 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.