A 24-Year-Old Woman With Intractable Seizures: Title and subTitle BreakReview of Surgery for Epilepsy

DR SHIP: Ms H is a 24-year-old woman with intractable seizures. Her seizures began at age 10 years and have been essentially unchanged since that time. They occur overwhelmingly at night, usually within an hour of sleep onset. They awaken her with a brief feeling of heaviness and movement in the right arm. She then develops rigid posturing of the right arm and hand and rotates her body toward the right. The left arm and hand will often reach out and grab the right arm to help control its movement. The right arm becomes intensely rigid for 30 to 45 seconds. When it is over, she feels exhausted and falls asleep. On occasion, she goes into a more prolonged, generalized tonic seizure that may last for 2 to 3 minutes. While her seizures have waxed and waned over the years, she usually has 2 to 10 of these seizures per night.

She has no history of brain injury or infection that might have predisposed her to seizures. While both sides of her family have a history of multiple types of seizures, neither parent nor her brother, her only sibling, has ever experienced a seizure. No affected relatives have seizures similar to hers.

Over the past 14 years, Ms H has been prescribed phenobarbital, lamotrigine, carbamazepine, topiramate, tiagabine, clonazepam, and ethosuximide. All were taken under close medical supervision but with little to no long-term effect on the frequency or severity of her seizures. Her medications at the time of her evaluation were oxcarbazepine, 300 mg in the morning and 750 mg at night; zonisamide, 200 mg in the morning and 400 mg at night; levetiracetam, 750 mg in the morning and 1500 mg at night; and lorazepam, 2 mg at night. She is allergic to lamotrigine, which causes a rash.

Ms H completed college and worked until recently as a preschool teacher. She has never smoked or used recreational drugs and drinks rarely.

On physical examination, Ms H was alert and oriented. She was tall, thin, and symmetrically developed, with no evidence for neurocutaneous disorders. Sense of smell was depressed in the right nostril. Her pupils were normal, as were her visual fields and lower cranial nerves. She was more coordinated with her left arm and forearm despite being right-handed. She had a negative Romberg but had decreased right arm swing when she walked, mild difficulty with right-hand rapid alternating movements, and postured the right hand when distracted. Detailed sensory examination was normal, as were her reflexes.

Ms H's laboratory measurements were notable only for a mild, normocytic anemia, with a hematocrit of 33%.

Her diagnostic studies included (1) a routine electroencephalogram (EEG) (Figure 1A) that showed left temporal and left central slow wave activity suggestive of localized cortical injury and rare left temporal interictal epileptiform discharges suggestive of an area of cortical irritability¹ ; (2) a magnetoencephalogram (MEG), which fortunately captured a seizure at onset and showed low-voltage fast and ultrafast activity over the left anterior-mesial-frontal region just prior to movement-related artifacts; (3) inpatient EEG telemetry on reduced medication (Figure 1B), which captured multiple seizures that showed a bicentral high-amplitude spike and wave discharge, followed by bicentral attenuation of EEG activity for 3 to 4 seconds, followed by bicentral higher-amplitude rhythmic activity associated with the onset of her clinical behavior and compatible with a mesial frontal origin¹ ; radiology studies including (4) 1.5-T magnetic resonance imaging (MRI), which was normal, (5) 3.0-T MRI (Figure 2) to better show cortical anatomy, which found mild diffuse cerebellar atrophy, perhaps related to long-term use of phenytoin, and a slightly enlarged right temporal horn, suggestive of a relatively small right temporal lobe, (6) MR spectroscopy (MRS) to assess the chemical composition of areas of the brain, performed in both temporal lobes and both mesial frontal regions, which showed slight decreases in the N-acetylaspartate concentrations in the left mesial temporal region, suggestive of loss of neuronal elements, and (7) positron emission tomography (PET) scan (Figure 3) when she was not having symptoms, showing a decrease in right temporal lobe use of fluorodeoxyglucose, also suggestive of decreased metabolic activity in that region; and (8) detailed clinical neuropsychological testing,^{2 - 3} which demonstrated moderate difficulty with frontal lobe executive functioning (right greater than left).

My seizures started when I was 10 years old. We started a regimen of different medications. After that—more medications, but none of them really seemed to have the desired effect. I guess that's why right now I’m seeking surgery. My seizure activity has been more frequent than 1 to 2 at night. When I was taking different medications, I was having up to 4 or 5 a day.

Right now I have 1 or 2 seizures a night, and they are mainly in the right side of my body. Usually just my right arm is involved. I understand that they're going to place electrodes in and onto the surface of my brain to know if I am a surgical candidate. If I am a surgical candidate and surgery is performed, I could still need medicines, but I would hope that we could decrease my medicines if I had surgery.

I feel that my seizures have prevented me from doing outdoor activities. I like to be active and go hiking, and I feel restricted because I don't feel safe when I’m hiking. I stopped working because one night, I had 10 seizures. I realized that I just can't keep working and have this many seizures. Working with children, especially, it was just too dangerous.

Some people would say my seizures are well controlled. But it's hard to say that seizures are well controlled when you have to be the person who is living with them. And because I’m not able to support myself right now, it is very hard to say that the seizures are “well controlled.”

Briefly discuss the epidemiology and pathophysiology of Ms H's type of seizures. What is the long-term harm of uncontrolled or poorly controlled seizures? Which antiseizure medications are most effective for this type of seizure and what are the long-term complications? What is the evidence that surgery is more effective than medication in controlling this type of seizure? Does the location of the seizure focus make a difference in effectiveness of surgery? What are the indications for surgery? What are the long-term outcomes and adverse effects? What does the future hold for persons with medically refractory seizure disorders? What do you recommend for Ms H?

DR SCHOMER: Ms H has simple partial and secondarily generalized seizures that are, by clinical description, likely to be of left mesial frontal origin; ie, supplementary motor cortex region. These seizures appear to be nonlesional and medically refractory. As shown by longitudinal measures of quality of life,⁴ patients like Ms H, with a history of years of medically refractory partial and/or generalized epilepsy who continue to have poorly controlled seizures, are subject to undereducation,^{5 - 7} underemployment,⁵ higher than expected coincidence of depression,⁸ paranoia and other severe and disabling psychiatric and personality disorders,^{7 ,9 - 10} disturbed memory, faulty cognitive functioning,^{11 - 13} and a shortened life expectancy, regardless of their baseline socioeconomic background.^{14 - 20}

Knowing these possible outcomes, many physicians have adopted an aggressive approach to treatment. While researchers are investigating the variables that are behind the seemingly progressive nature of this disorder,²¹ barring repeated closed head injuries and the resultant intermittent breakdown of the blood-brain barrier,²² no other theory reasonably explains this observation. Aggressive medical treatment, in which the treating physician either succeeds in adjusting medical treatment until the patient is seizure-free over long periods of time,^{6 - 7 ,10} or, if unable to control the patient's seizures, refers the patient for surgical evaluation, is now considered standard of care.²³ The medical community is more effective at recognizing and treating the seizures; ie, the outward manifestation of epilepsy, than it is at preventing or arresting the epileptic process itself, making epilepsy surgery an important alternative to medical therapy alone to help the patient become seizure-free.^{24 - 27}

Epilepsy is a clinical diagnosis that is based most importantly on the history as well as the clinical/neurological examination. Characterization and localization is aided significantly by tests of neurophysiological function such as EEG or EEG telemetry and, more recently, MEG. The workup is also aided significantly by a variety of radiological studies, including anatomical imaging of the brain such as MRI, high-resolution MRI,²⁸ MRS,^{29 - 30} and classic computed tomography (CT)³¹ scanning, as well as by tests of dynamic functioning that include single-photon emission CT (SPECT)^{32 - 34} scans, PET^{35 - 39} scans, and functional MRI (fMRI). The goal of performing multiple studies is to more precisely define the seizure focus; there is no evidence to support one type of study over another, although conditions for a given patient may make some tests less optimal. For example, MEG is a new technique that measures very small variations in the magnetic fields that are created by the electrical activity of the brain. These fluctuations in magnetic field strengths are not affected by structures such as the skull and, hence, theoretically are better for localization techniques.⁴⁰ However, the equipment for this procedure is expensive and unable to record continuously for the long periods of time that are often required to record seizures.⁴¹

The term epilepsy implies “recurrent seizures.” These occur when a sudden and pathological period of neuronal hypersynchrony is associated with an alteration of clinical behavior. Epilepsy occurs in approximately 1% of the population, with peak onset at both ends of the age spectrum.⁴² Quiz Ref IDThis disorder may be genetic in origin and, therefore, associated with a positive family history of seizures, or it may be sporadic and associated with any number of coexistent brain disorders: congenital/developmental cortical anomalies; infectious disorders such as abscess, meningitis or encephalitis; brain tumors; or as a consequence of a stroke or head injury. It may also occur more frequently in a setting where a patient has a genetic predisposition to epilepsy, as manifested by a positive family history for epilepsy, and has an associated risk factor for partial seizures, as noted in Ms H.

The International League Against Epilepsy (ILAE) in 1981⁴³ defined 2 major subtypes of seizures. When seizures occur and both hemispheres of the brain are suddenly, synchronously symptomatic, these are referred to as generalized seizures. Such seizures include classic absences, its variations such as childhood- or juvenile-onset absence epilepsy syndromes, and juvenile myoclonic epilepsy. Included in this category are seizures in which the main behavior is motor. These include generalized tonic (rigid), tonic/clonic (first rigid then repetitive jerking), or clonic (repetitive jerking). These disorders tend to start relatively early in life, have a genetic predisposition, and are by and large easier to treat with currently available medications than the other type of epilepsy. Generalized seizures may also accompany some of the most severe degenerative diseases seen in neurological practices, and these may or may not be amenable to medical management.⁴⁴ If epilepsy comes from a specific area of the cortex; ie, focal onset, it is referred to as partial epilepsy by the ILAE.^{31 ,43}

Quiz Ref IDPartial seizures can be further subclassified based on whether consciousness is impaired or the patient goes on to have generalized major motor seizures. If consciousness is not impaired, it is referred to as a simple partial seizure. If consciousness is impaired, it is a complex partial seizure. If the seizure progresses to generalized motor seizure activity, it is referred to as a secondarily generalized seizure. Most “auras” are thought to be simple partial seizures. Since partial seizures imply a focal onset, clinicians look for underlying focal brain pathology. They look for “lesions” in partial epilepsy based on a patient's report of an aura or the earliest symptoms, which are linked to the seizure because that area of the brain is likely to be at or near the origin of the pathological physiology. Neurologists try to localize the electrical onset of a seizure through EEG telemetry techniques and to follow the physiological progression of the seizure from the beginning to the end to relate the EEG changes to the behavior^{32 ,45} ; ie, seizure semiology. This process defines the anatomical-physiological substrates that are necessary for that particular seizure.

Historically, the earliest features of Ms H's seizure were motor in nature. On occasion, she did report a heavy or numb feeling in the right side at the beginning of the seizure, but this was not a consistent part of her seizures. The most consistent behavior was tonic elevation of the right arm off of the bed, occurring in the middle of the night after she had fallen asleep. This behavior was then followed by tonic head turning consistently toward the right, with the left arm coming across and grabbing the right arm. She believed she was awake for most of the events. The peculiar behavior that she exhibited is referred to as “versive.” Seizures of a versive nature frequently come from the supplementary motor cortex in the mesial frontal region (Figure 4).^{46 - 47} This area lies in front of the mesial extension of the motor cortex as it is draped over the interhemispheric fissure. It is a motor programming area and, as such, gives rise to behaviors that involve both sides of the body. It has been noted in clinical correlation studies that the side of the body that is rigid at the beginning of the seizure is usually contralateral to the seizure onset focus.^{46 - 47} Therefore, Ms H's seizures, by clinical description, are likely of left mesial frontal origin^{46 - 47} or could originate from a “silent” cortical area that projects to the mesial frontal lobe. Seizures of the supplementary motor area are also recognized to occur primarily during drowsiness or as the patient falls asleep, and they tend to cluster.⁴⁶

Because Ms H is alert and can remember most of her seizures, they are categorized as simple partial seizures. Her seizures are medically refractory since she continues to have frequent seizures despite numerous trials with reasonable medications under supervision.

Ms H's clinical neurological examination⁴⁸ did not reveal any obvious motor deficits or reflex asymmetries that would suggest either a hemiplegia or hemiparesis. The examination did show signs of right arm dysfunction that included a decrease in the ability to do rapid alternating functions and the appearance of abnormal posturing while doing stressed tasks. These tasks are signs of motor programming difficulty rather than a problem with strength or tone. Therefore, her neurological deficits support a left mesial frontal origin. Ms H's multiple diagnostic studies did not find a consistent identifiable lesion.^{49 - 50} Her study results actually suggest multifocal problems, including both frontal and temporal lobes.

Epileptologists believe that while a single seizure is unlikely to damage the brain, repeated seizures over long periods of time are likely to be associated with change in brain functioning.^{9 ,12 ,21}

In a study that evaluated the 3-dimensional features of neurons removed as part of a classical temporal lobe surgical procedure for mesial temporal lobe sclerosis (MTS),⁵¹ the investigators demonstrated that more synaptic neuronal elements were lost and the 3-dimensional complexity of the neurons was simplified in direct relation to the total lifetime accumulated number of seizures and the distance they were away from the focus. Quiz Ref IDThese results support the view that changes in neuronal ultrastructure occur over long periods of time and are associated with a relatively large numbers of seizures. Therefore, an aggressive medical approach to treatment early in the care of patients with recurrent partial seizures seems justified to theoretically prevent such changes.

Approximately 10 medications currently available would be considered appropriate for Ms H.⁵² Of those 10, phenytoin,^{53 - 54} carbamazepine,⁵³ lamotrigine,^{55 - 56} and oxcarbazepine^{57 - 58} are considered first-line drugs; ie, effective as monotherapy based on controlled studies. Levetricetam,^{59 - 60} topiramate,^{61 - 62} valporate,^{63 - 64} and zonisamide⁶⁴ are all reasonable second-line drugs that can be added if the primary drug fails. It should be noted that no controlled trial shows that any of the primary drugs are better than the others when it comes to seizure control. Thus, physicians choose which drug to use initially based on a complex interplay of their experiences with these medications and their anticipated efficacy as well as long- and short-term adverse effects, the possible effects of the anticonvulsant on pregnancies or birth control, or possible drug-drug interactions, particularly in the case of older patients.⁶⁵

Expense, which can come from both the actual cost of the medication itself and the anticipated follow-up requirements, such as repeated hematologic or hepatic function testing, is considered but is usually not as critical as other factors, depending on the patient's insurance and ability to pay for the cost of medications. Once a decision about medication is made, most experts opt to use a single drug therapy.⁵²

Generally, the dose of the medication is gradually increased to a near toxic level before considering a second drug. A trial with a second drug is indicated if there is an allergy or adverse reaction to the first medication or if seizures continue. When changing to a second medication, the second drug is added and titrated to therapeutic levels before tapering and discontinuing the first drug. This “monotherapy” approach tends to minimize drug-related adverse effects, reduce the incidence of toxic effects, and improve adherence, and is associated with reduced pharmacokinetic and/or pharmacodynamic interactions.^{66 - 67}

Rational polypharmacy, however, is a reasonable approach when patients have shown that they are refractory to multiple single-drug trials. Clinicians try to choose drugs for which the pharmacokinetic and pharmacodynamic interactions are few and the 2 drugs have synergistic mechanisms of action. A study in 525 patients with epilepsy in Scotland suggested that about half of all patients with new-onset epilepsy have their seizures controlled with the first drug that is used,⁶⁸ regardless of choice. Adding a second drug leads to only 13% more of the population becoming seizure-free, and this, also, appears to be regardless of drug choice. A third drug results in freedom from seizures in only an additional 1% of the population. Therefore, about one-third of all patients with new-onset epilepsy will have seizures refractory to medication and becomes potential candidates for surgical evaluation. The study did not determine whether any specific seizure semiology is associated with becoming seizure-free.

Symptomatic partial epilepsies rarely spontaneously remit and, therefore, medical treatment is almost always for life,⁶⁹ and very little is known about the long-term consequences of anticonvulsant use, especially with respect to the newer drugs. Some of the older drugs are associated with an increased risk of certain lymphomas, the development of progressive atrophy of the cerebellum, and the development of a progressive peripheral neuropathy.⁷⁰ In children, some of the currently used drugs are associated with decreased learning above and beyond what might be expected from recurrent seizures.¹³ Finally, medications also reduce selective aspects of the immune system, increasing susceptibility to a variety of infections.⁷¹

The goal of surgery for epilepsy is to control or eliminate seizures and reduce the need for antiseizure medications. In the only prospective trial that compared continued “best medical” treatment with surgery,⁷² those who underwent surgery had better seizure control (58% of those randomized to surgery and 64% of those who actually underwent surgery were seizure-free at 1 year) and better longer-term quality-of-life measures. Quiz Ref IDHowever, surgical success depends predominantly on the identification of an epileptogenic or seizure onset zone, so surgical evaluation aims to identify patients who have clear focal and reproducible onset of seizures. Data from neuropsychology testing^{2 - 3 ,73} and/or imaging studies³¹ should support the same region as the focus seizure onset. Additionally, the region, or zone, in which the focus is found, when removed, should not lead to significant additional neurological, cognitive, or psychiatric dysfunction.⁷⁴ The most common surgical approaches to medically refractory epilepsy in a patient with MTS are mesial temporal lobe removal or amygdalohippocampectomy, with or without a partial anterior temporal lobectomy.⁷⁵

Surgical success is measured by either the Engel classification⁷⁶ or the ILAE classification⁷⁷ methods (Box). In both of these classification schemes, success is measured by relative seizure freedom for 1 year or more postoperatively. Neither of these classification schemes measures other factors such as quality-of-life measures.^{4 ,78 - 79} However, most groups that report quality-of-life outcomes find a direct correlation between quality of life and freedom from seizures.⁸⁰

For patients with MTS, surgical success, as defined by Engel classes I and II or ILAE classes 1, 1a, and 2,^{25 ,75} (Box) is greater than 90% in most US medical centers.⁷⁵ About 45% of patients become seizure-free (Engel class I or ILAE class 1 or 1a) (Box) and up to half can stop medication, while 45% have auras^{75 - 77} (ILAE class 2) or infrequent seizures and require medication (ILAE class 3). From 8% to 10% are no better (ILAE class 4 or 5) or worse (ILAE class 6) than before surgery. Patients with other temporal lobe lesions (depending on the underlying pathology) also generally do better than do patients without an identifiable lesion The same observation has been made in cases of frontal lobe epilepsy; however, far fewer studies of frontal lobe–onset seizure surgery have been conducted (Table). Overall, the surgical outcomes in frontal lobe–onset seizures are not as good as for temporal lobe–onset seizures in terms of seizure freedom varying from 54% to 66% (Engel class I or ILAE classification 1 or 1a).^{75 ,88 - 90} If there is no lesion, the Engel class I or ILAE class 1 or 1a surgical outcomes at 1 year vary from 29% to 41%.⁷⁸

DR BLACK: Neurosurgeons have an important role in both the diagnosis and treatment of patients with epilepsy and are an integral part of the comprehensive epilepsy management team. In helping to identify a seizure focus, they assist the epileptologist in deciding when to use invasive recording techniques and which techniques to use. In assessing the site of seizure onset, the surgeon may use invasive monitoring, including depth electrodes^{91 - 92} ; subdural grids and strips of electrodes⁹³ ; epidural pegs⁹⁴ ; or foramen ovale electrodes.⁹⁴

Each of these electrode types has its particular indications, risks, and benefits. For example, subdural grids and strips do not penetrate the brain but allow stimulation of motor, sensory, or speech activity as well as recording,⁹⁵ but, they may not accurately localize a deep focus as well as a depth electrode.^{91 ,93} In Ms H's case, both depth electrodes and strips were used and were successful at demonstrating a likely electrical seizure source.

Neurosurgery aims to control medically intractable⁷² seizures. If 3 successive anticonvulsants do not control seizures, surgery should be considered.⁶⁸ Many procedures can be used to treat seizures, but the most successful remove a seizure focus and its related epileptogenic zone, which has been identified by both imaging and electrophysiology.⁹⁶

Examples of these foci include focal cortical dysplasia, hamartoma, or other congenital lesions; a brain tumor; or brain sclerosis including MTS. Sophisticated imaging techniques with 3.0-T MRI may show lesions not recognized in routine imaging.²⁸

For anatomic abnormalities other than MTS, new image-guided techniques, which include intraoperative imaging, have made very precise resection and verification of resection possible. Quiz Ref IDThe amount of tissue removed may vary from lesionectomy (removing just the visualized lesion) to topectomy (removing an electrophysiological focus) to partial lobectomy (removal of a portion of the frontal, temporal, or, rarely, occipital lobe) or multilobar resection. The most striking example of a resection is a hemispherectomy,⁹⁷ usually performed in young children with multiple intractable seizure foci localized to one hemisphere. Patients reported in the literature often do very well, becoming seizure-free or with improvement rates approaching 90%.⁹⁷

Patients with lesions, regardless of their location in the brain, generally have better seizure outcomes than do patients without identifiable lesions,^{78 ,96 ,98} who have tissue removal based on electrophysiological data alone. In general, extratemporal sites have a success rate (Engel class I or II) of about 80%.^{25 ,50 ,75}

For any brain site, the risk-benefit ratio must be considered. Another surgical approach is to divide the fibers that conduct seizures between cortical columns; ie, a subpial transection.⁹⁹ If seizures are emanating from motor or sensory or other eloquent cortex, brain tissue that cannot be removed without causing permanent neurological changes, it may be possible to prevent seizure spread by multiple trenches placed in the cortex that prevent lateral spread but do not affect the transmission of impulses to the deep structures.⁹⁹ These patients seldom have a sensory loss or motor weakness after the procedure. Such patients have a 40% to 50% chance of improving seizures; ie, reducing seizure frequency by more than 50%.⁹⁹

Another example of dividing fibers is the corpus callosotomy,^{100 - 101} in which dividing all or part of the corpus callosum prevents rapid transmission of seizure impulses from one hemisphere to another. It has an important role in reducing seizures with “drop” attacks.^{100 - 101}

Finally, vagal nerve stimulation^{102 - 103} has recently emerged as a technique to treat seizures originating from multiple cortical sites. This procedure involves placing electrodes around the left vagus nerve in the neck and attaching them to a stimulator that is placed under the clavicle or in the axilla. The stimulator is programmed by an external device. It is believed that antidromic stimulation of the vagus lessens seizure potential in a distributed network.¹⁰² Vagal nerve stimulation is sometimes used when a focus cannot be clearly defined or when there are multiple foci present.^{102 - 103} Vagal nerve stimulation has a 45% to 50% improvement rate; ie, greater than 50% reduction in seizure frequency, but is attractive because it does not involve removal of brain tissue or invasion of the brain.¹⁰⁴

This may be a reasonable approach for Ms H if her seizures are coming from eloquent cortex or she is thought not to be a surgical candidate for other reasons. The long-term consequences of epilepsy surgery depend on 2 variables, control of seizures and location (type) of surgery. If the seizures are completely controlled and the patient is able to discontinue antiseizure medication, quality-of-life measures improve significantly and the consequence of the surgery is principally dependent on the location and type of surgery.¹⁰⁵

For example, if patients have removal of a partially functional hippocampus in the language-dominant hemisphere, they may experience a decrease in memory function related to that area. Similarly, removal of any partially functional area may lead to either temporary or permanent loss of function. However, if the surgery is not helpful in gaining some element of control, the patient may continue to have all of the potential problems related to long-term medication use but also have additional superimposed surgically related deficits.

Ms H had a combination of strips and depth electrodes placed for recording purposes. The depth electrodes were directed from a lateral approach toward both mesial temporal lobe regions (hippocampi) because of the findings on routine EEG and MRI/MRS. She also had multiple strips placed through burr holes along the high parasagittal midline region to position their recording electrode contacts along the mesial (interhemispheric) fissure where the supplementary motor cortex regions reside. She had multiple seizures with seizure onset localized to the more posterior aspects of the supplementary motor cortex region on the left. This area was believed to be the cortical region for foot and lower leg function.

Ms H has frontal lobe−onset seizures likely originating in the posterior-inferior mesial frontal region but without an associated anatomical lesion. Seizures of the frontal lobe are frequently divided into anatomical regions because of the varying clinical presentations of the different regions. At least 8 regions of interest have been identified, including the primary motor cortex, insula, mesial frontal, frontopolar, orbital frontal, cingulate gyrus, lateral frontal convexity, and central regions (Figure 4). While surgical outcome is still maximally dependent on presence or absence of a lesion,^{106 - 107} specific neurological and cognitive/behavioral deficits relate to these anatomical regions (Figure 4).

Outcomes from case series of frontal lobe surgery (Table) show minor variability depending on the study methods and when it was conducted. However, the outcomes are still best correlated with the ability to demonstrate a lesion using various imaging technologies. In addition, these studies demonstrate the difficulty of making an accurate prediction of outcome for a specific patient. It is difficult to know which data are most relevant to an individual's situation. For Ms H, the focus in the posterior aspect of the supplementary motor cortex creates 2 potential problems for the surgeon: first, there is no anatomic lesion, so this becomes a “nonlesional extratemporal” focus, which has a 29% to 56% chance of being helped by resection, ILAE class 1a^{81 - 83 ,106 - 107} ; and second, resection of this area can lead to possible permanent, though more likely temporary, loss of function in the right foot/leg.

DR SCHOMER: There are no data that directly answer the question of Ms H's likely surgical outcome because she has no lesion and has seizures originating in the posterior-inferior mesial frontal region, a quasi-eloquent region. Removal of that area might be associated with temporary or permanent motor deficits in the right foot or distal leg and/or a temporary Broca-like aphasia. These outcomes could be better understood following an fMRI study of right foot and leg functioning to determine if there is a blood flow response to motor or sensory functioning of the leg and foot or around the region in question. Also, an intracarotid sodium amobarbitol test for language and memory lateralization will be predictive of the likely temporary Broca-type aphasia.³ The seizure-free percentages from surgery on this area of the brain are, at best, 55% to 60% (Table).

However, MRI technology and other methods of brain functional investigation are changing rapidly. Ms H might benefit from a vagal nerve stimulator placement^{102 - 104} to give her some relief. The preoperative surgical data including the EEG data from the invasive electrodes will continue to be useful for quite some time. Thus, because the data are confounding, I would opt to use vagal nerve stimulation, follow her progress closely for another few years, and make a definitive decision within that time frame, perhaps bringing to bear improved technologies.

DR BLACK: It is reasonable to proceed with vagal nerve stimulation, but I doubt that it will provide major or long-lasting relief. The most important decision maker in this setting is the patient. The surgeon's role is to present the options and likely outcomes. As an alternative to vagal nerve stimulation, or if it fails, I would recommend cortical surgery using intraoperative brain mapping and corticography to define the limits of the seizure focus and resect it. This provides an opportunity to identify and remove the source of seizures in a definitive way. If these studies suggest that removal will produce unacceptable hemiparesis, I would perform subpial transections.

Partial epilepsy implies a localized-onset seizure and, hence, a likely localized lesion. This form of epilepsy accounts for approximately 60% of all epilepsies and occurs in about 1% of the population. Best medical treatment will control about 50% to 65% of patients. Fully one-third will have inadequate control with currently available medications. From this group of nonresponders, clinicians select patients to undergo presurgical evaluations to determine if their seizures are coming from a consistent focus and if that area of the brain can be removed without leading to additional disabling neurological deficits. If this is possible, patients further undergo evaluations to determine if they can psychologically deal with this type of elective brain surgery, since active psychiatric illness could complicate the postoperative course.⁷⁴ In most closely watched series of postoperative patients, patients who have a convergence of preoperative data and identifiable radiological lesions have the best outcomes.

QUESTION: What are the risks for death in a case like this?

DR SCHOMER: Sudden, unexpected death in patients who have medically refractory epilepsy occurs 1 in 250 patient-years.^{15 ,18} While that may not seem like a significant number, this young woman, who started having seizures when she was 10, by the time she reaches 70 or 80 is going to have accumulated 60 or 70 years of uncontrolled seizure activity. Her overall risk of sudden, unexpected death thus becomes more significant.

Perhaps even slightly greater is the risk that the patient could do harm to herself during a seizure; eg, falling down a flight of stairs or aspirating. Patients have been known to drown while taking a bath, burn themselves while cooking, or walk out into oncoming traffic while having a seizure.

QUESTION: Did Ms H's relatives have nocturnal frontal lobe seizures? And if the patient does have autosomal dominant frontal lobe epilepsy, what do you know about surgical results in these patients?

DR SCHOMER: We know virtually nothing about surgical outcomes in those cases. It was difficult to clarify her family history of seizures. While many people were identified with seizures, it was difficult to identify what their seizure behaviors were. Her parents did say that her seizures seemed to be unique in terms of their presentations compared with the rest of the affected members. I think she has a genetic-based predisposition to seizures but does not have the genetically dominant nocturnal frontal lobe seizures related to a specific nicotine receptor disturbance.⁸⁴

QUESTION: The likelihood of seizure control after 2 or 3 medication trials is disappointingly low.⁶⁸ Given that going for 10 or 12 years with medically intractable seizures is not that unusual before the first presurgical workup is undertaken, what are your recommendations regarding the timing of a presurgical evaluation?

DR BLACK: As adult epileptologists, we often don't see patients until they are teenagers or in their early 20s. I think pediatric epileptologists would probably agree that an early aggressive approach to focal epilepsy is probably better because of the ability of the brain to make adjustments.

QUESTION: Although Ms H was on 4 medications, it is not clear that any of them were really doing much good. At what point, even if the vagal nerve stimulator were not effective, would you consider withdrawing at least 1 or 2 of them?

DR SCHOMER: The list of medicines that you saw was the list of medicines that she presented with. Since then, her regimen had been reduced to 3 drugs, and 2 of those were new drugs. So we had completely changed her drug approach. We hope that we eventually get rid of at least 1 or 2 of the more sedative ones.