Original Article

Surgical Outcomes for Intractable Epilepsy in Children With Epileptic Spasms

Journal of Child Neurology 27(6) 713-720 ª The Author(s) 2012 Reprints and permission: sagepub.com/journalsPermissions.nav DOI: 10.1177/0883073811424463 http://jcn.sagepub.com

Brian D. Moseley, MD1, Katherine Nickels, MD2, and Elaine C. Wirrell, MD2

Abstract Epileptic spasms, or seizures marked by flexor, extensor, or flexor-extensor spasms, are not always responsive to medical management. The purpose of our study was to evaluate the outcome of epilepsy surgery in children with medically intractable epileptic spasms. We identified 11 children with epileptic spasms who underwent lesionectomy (36%), lobectomy (27%), multi-lobectomy (9%), hemispherectomy (18%), or corpus callosotomy (9%). At the time of surgery, 6 children had developed other concurrent seizure type(s), including simple partial (9%), complex partial (27%), partial undifferentiated (9%), primary generalized tonic clonic (9%), tonic (9%), atonic (27%), and myoclonic (9%) seizures. Six children (55%) were seizure free at last follow-up from initial surgery. Predictors of favorable outcome included lack of focal slowing and the presence of less than 2 interictal epileptiform abnormalities on postoperative electroencephalogram (P ¼ .035 and .035, respectively). Favorable outcome was significantly associated with parent/caregiver report of improved postoperative developmental outcomes (P ¼ .026). Keywords epileptic spasms, infantile spasms, epilepsy surgery Received May 11, 2011. Received revised September 1, 2011. Accepted for publication September 1, 2011.

The treatment of epilepsy in children may be complex and can require more than pharmacologic intervention. It is estimated that 10% to 40% of children will continue to suffer from disabling seizures despite adequate antiepileptic drug management.1–5 For some children with medically intractable focal onset seizures, surgical intervention has been proven to be efficacious.6–14 However, the utility of surgical interventions in children with seizure subtypes such as epileptic spasms has been less robustly explored. Although medical management— including adrenocorticotropic hormone, corticosteroids, antiepileptic drugs such as vigabatrin and valproic acid, and the ketogenic diet—can be effective at eliminating these seizures, it is sometimes not enough.15 Investigation into surgical treatment is warranted, given the unfavorable outcomes experienced by many children with medically intractable epileptic spasms. A previous longitudinal study by Riikonen demonstrated that despite standard medical treatment, the majority of these children continued to suffer from seizures (58%) and intellectual impairment (76%) decades after diagnosis. Mortality was also increased, with 31% of children dying within the first 30 years following diagnosis. Almost one-third of such deaths were prior to the age of 3 years, while 61% were prior to the age of 10 years.16 Some cases of epileptic spasms lack a discernable cause. These cases are described as either idiopathic (provided the child demonstrates normal development prior to the onset of

spasms and no other neurologic sequelae) or cryptogenic (the child’s development is impaired and/or suffers from other neurologic sequelae). However, the majority of epileptic spasms are symptomatic, or secondary to clearly defined underlying causes.17 These can include hypoxic-ischemic encephalopathy, trauma, encephalitis, meningitis, neonatal hypoglycemia, metabolic disorders (including pyridoxine dependency, phenylketonuria, maple syrup urine disease, biotinidase deficiency, Menkes disease, Leigh disease, Krabbe disease, non-ketotic hyperglycinuria, and hyperammonemia), and genetic epileptic encephalopathies (including CDKL5 and ARX).17 There are also structural lesions (including malformations of cortical development, tuberous sclerosis, tumors, arteriovenous malformations, and cerebral abscesses), which can result in epileptic spasms and potentially be amenable to surgical resection.17 The purpose of our study was to evaluate the outcomes of children with medically intractable epileptic

1 2

Department of Neurology, Mayo Clinic, Rochester, MN, USA Divisions of Child and Adolescent Neurology and Epilepsy, Department of Neurology, Mayo Clinic, Rochester, MN, USA

Corresponding Author: Brian D. Moseley, MD, 200 First Street SW, Rochester, MN 55905 Email: [email protected]

714 spasms who underwent epilepsy surgery at our institution. In addition to assessing for seizure freedom, we evaluated for potential historical, electrodiagnostic, and radiographic features that portended a better postoperative prognosis.

Materials and Methods The medical records of all children aged 0 to 18 years undergoing epilepsy surgery, including lesionectomies, lobectomies, multilobectomies, hemispherectomies, and/or corpus callosotomies, at Mayo Clinic in Rochester between January 2002 and June 2006 were reviewed. Children who underwent implantations of vagal nerve stimulators alone were excluded. All children who had epileptic spasms at the time of surgery were selected for study. Epileptic spasms were defined as flexor, extensor, or flexor-extensor spasms. At least 1 preoperative electroencephalogram (EEG) had to demonstrate hypsarrhythmia or modified hypsarrhythmia. Hypsarrhythmia was defined as a disorganized background with multifocal spikes, high-amplitude slowing, and spasms marked by high-amplitude slow transients followed by a generalized electrodecrement.17 Modified hypsarrhythmia included asymmetrical hypsarrhythmia, hypsarrhythmia with a single focus, hypsarrhythmia with increased interhemispheric synchronization, hypsarrhythmia with episodes of attenuation, and hypsarrhythmia with predominately highvoltage slow activity and little sharp-wave or spike activity.18 Although children were screened for developmental delay/regression, this was not essential for study inclusion. Given that children who developed spasms beyond the age of 12 months were eligible for inclusion, we did not use the term infantile spasms to describe the seizures. In children meeting the above criteria, clinical and electrophysiological data were reviewed by the authors. Data acquired included gender, age at afebrile seizure onset, preoperative neurologic examination abnormalities, number of previous antiepileptic drugs tried, previous use of vigabatrin, previous use of corticosteroids and/or adrenocorticotropic hormone (ACTH), and previous use of the ketogenic diet. The presence of other seizure types (simple partial, complex partial, partial undifferentiated, secondarily generalized tonic clonic, primary generalized tonic clonic, tonic, atonic, and/or myoclonic) prior to surgical intervention were recorded. The details of presurgical imaging, including brain magnetic resonance imaging (MRI), subtraction ictal single-photon emission computed tomography (SPECT) coregistered to MRI (SISCOM), and/or fluorodeoxyglucose positron emission tomography (FDG PET) were examined. Prolonged EEG reports obtained immediately prior to surgery were reviewed; the presence/ absence and location of interictal and ictal abnormalities were recorded. The details of each child’s surgical intervention were recorded, including age at surgery, type(s) of surgery performed, pathologic findings in resected tissue, completeness of resection of the radiological abnormalities present on preoperative imaging, and surgical complications. If a child underwent multiple surgical resections, the details of the first surgery were thoroughly examined. Outcomes were assessed at most recent follow-up at our institution using the modified Engel classification.19 For those children having more than 1 surgery, outcomes were assessed following the first surgery (immediately prior to the next surgical intervention) and at most recent follow-up after the final surgery. Favorable seizure outcome was defined as an Engel classification of I-II and unfavorable outcome as Engel classifications of III-IV. Developmental quotients (defined as the child’s functional/ developmental age divided by the chronological age  100) were calculated prior to surgery and at last follow-up, based on developmental milestones recorded in the medical record and/or formal

Journal of Child Neurology 27(6) neuropsychometric testing. This allowed us to assess for developmental changes following surgery. All developmental quotients less than 20 were grouped into a single category (