University of Virginia Health System

The Neuro Center's physician newsletter. Click here to go to Neurogram's home page or the University of Virginia Health System's Epilepsy pages.

Recent Advances in the Treatment of Epilepsy

Epilepsy affects an estimated 2 to 4 million people in the United States. Although it can have devastating consequences to those affected, it need only cause minor inconveniences with proper management. In the last several years, recent additions to treatment strategies include newly approved antiepileptic drugs (AED), vagus nerve stimulation, and epilepsy surgery.

ANTIEPILEPTIC DRUGS
The use of AEDs continues to be the mainstay of epilepsy therapy. The conventional AEDs, including phenytoin (Dilantin), carbamazepine (Tegretol), valproate (Depakote) and phenobarbital, remain the most widely prescribed AEDs. AED use has historically been problematic because of the narrow therapeutic window of dosing, frequent dose-dependent CNS side effects and rare but severe idiosyncratic end-organ toxicities. Therefore, there has been an active search for AEDs which are better tolerated, safer, and more effective.

A new era of pharmacotherapy for seizures was begun in 1993 with the approval of felbamate (Felbatol) which was the first new AED approved since the approval of valproate in 1978. The initial enthusiasm for felbamate was aborted when 35 cases of aplastic anemia and 18 cases of hepatitis were reported after it had been taken by approximately 100,000 patients. This has relegated it to use by epileptologists for refractory epilepsy where the potential benefits outweigh the risks. However, seven additional AEDs have been FDA-approved in a steady stream since then including gabapentin (Neurontin), lamotrigine (Lamictal), tiagabine (Gabitril), topiramate (Topamax), levetiracetam (Keppra), oxcarbazepine (Trileptal), and zonisamide (Zonegran).

The pharmacopeia of AEDs now available makes choosing among them difficult. There are several general principles of contemporary AED use which apply to new, as well as conventional, AEDs.
1) Select an AED based on seizure type and epilepsy syndrome.
2) For focal-onset seizures, select an AED based on pharmacokinetics, side effect profile, dosing frequency, and cost since all available AEDs (except ethosuximide) are efficacious.
3) Start one AED at a time.
4) Start AEDs at a low dose and increase gradually. The dose escalation recommended in the labeling for AEDs may be well tolerated by otherwise healthy patients, but a slower dose escalation is usually necessary for patients on concomitant AEDs or psychoactive drugs, or who have concomitant illnesses.
5) Increase the dose until either it is effective or side effects occur to define the maximum tolerated dose before deciding that it is ineffective.

FOCAL ONSET SEIZURES
Partial (or focal) onset seizures are the most common seizure type in adults. Consequently, all of the conventional and new AEDs listed above have been studied in this patient population and are approved as adjunctive treatment for focal-onset seizures. No head-to-head comparisons of the new AEDs have been performed, but it appears that all are equally efficacious for focal seizures; they reduce the seizure frequency by >50% in 30-40% of patients when added to patients' existing AEDs. While this demonstrates a lack of overwhelming efficacy of any one AED, it also illustrates that some patients who do not respond to one AED, will respond to another. The factors predicting who will respond are unknown.

All AEDs studied so far which have efficacy as adjunctive therapy for focal-onset seizures are also effective as initial monotherapy for focal onset seizures and some are FDA approved for this use. However, most physicians continue to start with a conventional drug as the first AED because of familiarity, known side effect profile, and reduced cost. It is likely that as the "new" AEDs become more widely used with more patient exposures to demonstrate safety, that they will replace conventional AEDs as initial monotherapy.

GENERALIZED SEIZURES
Generalized seizures are the most common class of seizures in children and include absence (previously termed petit mal), primary generalized tonic-clonic (previously termed grand mal), myoclonic, and atonic (drop attack) seizures. Absence seizures are generally known to be very responsive to valproate and ethosuximide, which is a conventional AED that is useful only for absence seizures. Among the new AEDs, lamotrigine is effective but the efficacy of most others has not been systematically studied. Lennox Gastaut syndrome is a particularly refractory form of epilepsy with a mixture of seizure types, mental retardation and characteristic EEG changes. Valproate is the conventional AED that has been most effective in this case, but lamotrigine and topiramate are also effective. The benefit of felbamate often outweighs the risks for patients with Lennox Gastaut syndrome.

THE NEWEST DRUGS
The most recently approved new AEDs are levetiracetam, oxcarbazepine and zonisamide. These drugs have many features in common with other new AEDs. All are indicated as adjunctive therapy of partial seizures in adults, and oxcarbazepine is also indicated as monotherapy. There is also evidence that they are effective for generalized tonic-clonic seizures but their efficacy for other seizure types is unknown. Their mechanism of action is not completely known. They are generally well tolerated by can cause somnolence and ataxia. They are not associated with serious hepatic, hematologic or any other end-organ toxicities. They are not heavily protein bound and have relatively little influence on other AEDs. Drug levels are available but of unknown value.

Levetiracetam (Keppra) was FDA approved in 1999. The University of Virginia is participating in pharmacokinetic and efficacy studies in children that suggest it is effective, but it is not approved for use in children. Dosing is started at 250 mg or 500 mg daily (10-20 mg/kg in children), and is increased weekly in 500 mg increments to a target dose of 3000 mg (60 mg/kg) divided BID, but up to 6000 mg/day is sometimes necessary.

Oxcarbazepine (Trileptal) was FDA approved in 2000 for the treatment of children 4-16 years old as well as adults. The drug is a chemical "cousin" of carbamazepine, having the same basic structure with an additional keto group attached to the main ring structure. This alters the metabolism so that it is reduced to the monohydroxy derivative (MHD) rather than the 10,11 epoxide metabolite of carbamazepine which is primarily responsible for the side effects of carbamazepine. Dosing is started at 600mg/day (8-10 mg/kg) divided BID with increases of 300 mg/day weekly to a target dose of 1200mg/day.

Oxcarbazepine is generally well tolerated with fewer side effects than carbamazepine. Approximately 30% of patients who have had sensitivity reactions to carbamazepine will also be sensitive to oxcarbazepine. No laboratory monitoring is required, although hyponatremia occurs in about 3% of cases.

Zonisamide (Zonegran) has been used extensively in Japan for a decade and was FDA approved in 2000. It is also being used in children in Japan, and there is evidence that it is effective for generalized tonic-clonic and myoclonic seizures and infantile spasms. Dosing is started at 100 mg/day with increases of 100 mg/day every two weeks to a target dose of 400 mg/day divided BID.

During studies of zonisamide in the US and Europe, approximately 2.5% of patients developed renal calculi, possibly because zonisamide is a weak carbonic anhydrase inhibitor. Therefore, adequate hydration is important. An uncommon peculiar side effect in children is oligohidrosis (reduced sweating).

Patients with refractory seizures, and some with new-onset seizures, are candidates for experimental AED drug studies. Pregabalin and harkoseride are such drugs that have demonstrated promise in studies at UVa and elsewhere.

VAGUS NERVE STIMULATION
Electrical stimulation of the left vagus nerve was FDA approved in 1997 for adjunctive therapy of partial epilepsy in adults and children over 12 years old. Controlled trials using low (placebo) versus high intensity stimulation demonstrated an ~ 50% reduction in seizure frequency in ~ 30% of those using the device for three months and ~ 50% after 18 months.

The exact mechanism of action is not known but the vagus nerve projects to brainstem nuclei which may have an effect on seizure suppression.

A signal generator is implanted in the upper left chest and is connected to a bipolar stimulation lead which attaches to the left vagus nerve in the neck. A handheld wand connected to a laptop computer is placed over the chest to program or interrogate the device. After the device is implanted, the output current and on/off time are adjusted upwards as tolerated and as needed for seizure control. By using accompanying magnets, patients or their family members can turn on the stimulator to attempt to abort a seizure or to lessen its duration.

A significant drawback is that the electrode wires are not removed even if the device is ineffective, because removing them is likely to cause more harm than leaving them in place. The generator may or may not be removed depending on the situation but the surgical scar, although small, will remain.

Side effects associated with each stimulation include mild hoarseness, coughing, and dyspnea (e.g. for 30 sec. every 5 minutes). Patients usually habituate to this surprisingly well.

EPILEPSY SURGERY
Epilepsy surgery has traditionally been viewed as a treatment of last resort but contemporary methods make it safe, effective and the only true cure for epilepsy. It should be considered early in the patient's course to minimize the consequences of poorly controlled epilepsy.

Surgery should be considered when the patient has failed adequate monotherapy trials with two appropriately chosen anticonvulsants since the chance of achieving adequate seizure control with additional AEDs is <20%. In order to determine medical intractability, non-compliance with medications must be excluded as a cause of poor seizure control and it must be verified that either the frequency of seizures and/or medication side effects are sufficient to cause a significant impairment in quality of life.

TESTING
After determining that the patient could potentially benefit from surgery, a diagnostic evaluation must be performed to determine actual suitability for surgery. This evaluation includes neuroimaging, inpatient intensive video/EEG monitoring, and neuropsychologic testing.

Magnetic resonance imaging of the brain with specially designed high resolution sequences are essential to identify structural abnormalities which can be the source of seizures. Examples of this include asymmetric atrophy of the temporal lobe or hippocampus, hippocampal sclerosis, cortical dysplasias, vascular malformations, and tumors.

Single photon emission computed tomography (SPECT) aids in localization by identifying areas of altered perfusion in the suspected epileptic focus. An interictal SPECT scan will either show no asymmetry of perfusion or decreased perfusion in the epileptic zone, presumably due to decreased neuronal activity interictally. Ictal SPECT is obtained by injecting radionuclide tracer during a seizure or within 90 seconds of seizure onset. The site of seizure onset will demonstrate increased perfusion representative of an area of hyperactivity. This requires knowing the exact time of seizure onset and availability of the radionuclide, which requires inpatient video/EEG monitoring. Positron emission tomography (PET) may demonstrate focal hypometabolism in the interictal state.

Inpatient intensive telemetry EEG monitoring with simultaneous video recording is performed to capture seizures. The goal is to confirm that the spells of interest are seizures, that they arise from a single location and to further identify the exact electrographic origin. Traditionally, at least three to four typical seizures are captured, although often more seizures are needed and rarely only one or two seizures is sufficient.

Should a definite origination zone not be adequately identified by these methods, EEG recording from intracranial electrodes may be necessary for adequate focus localization. The initial evaluation leads to a region suspicious for seizure onset and guides electrode placement. Electrodes can be placed into the brain substance (e.g., depth electrodes placed into the temporal lobes) or onto the surface of the brain (e.g., subdural grids and strips). The electrodes can remain in place for several days in order to capture seizures and localize their onset.

Neuropsychologic testing is performed on all surgical candidates to aid in lateralization of language and memory and to help predict the risk of functional loss after a surgery. When considering temporal lobectomy, the intracranial Amytal test (or Wada test) is performed in order to identify hemispheric language dominance and whether the contralateral temporal lobe is adequate to support memory function after surgery.

SURGERY TYPES
Many different types of surgery are now performed for refractory epilepsy, including anterior temporal lobectomy, cortectomy, lobectomy, lesionectomy, hemispherectomy, corpus callosotomy, and multiple sub-pial resections.

Anterior temporal lobectomy is the most commonly performed resective epilepsy surgery because "mesial temporal sclerosis" is the most common single identifiable pathology associated with epilepsy. The anterior 3 to 3.5 cm of the temporal lobe and mesial temporal structures are resected. The only randomized controlled trial of surgery versus medical therapy found 64% of those who underwent surgery were seizure-free versus only 8% of those who had maximal medical therapy. In most series and at UVa, approximately 70% of patients treated with anterior temporal lobectomy become essentially seizure free, although they may continue to have auras. Medications are routinely continued for at least a one year seizure-free interval prior to any consideration of withdrawal. There is only a 1% risk of serious complication, such as stroke. There is a 10-15% risk of "annoying" transient complications such as headache, depression, and speech or memory difficulties; rarely patients have noticeable memory impairment or visual field cuts.

Extratemporal resections (resection of regions other than the temporal lobe) are slightly less effective as only approximately 50% are seizure-free. However, this compares vary favorably to the natural history of this type of epilepsy, in which less than 10% of patients become seizure-free through AED use. The complication rate for extratemporal resection is also low, but the potential for functional deficits is greater when the seizures arise in important brain areas. Seizures arising in important brain areas, such as the frontal lobe near the motor cortex or language region of the dominant hemisphere, require careful mapping of the important regions to avoid including them in the resection. Mapping is most commonly accomplished by electrical stimulation of the electrodes to identify the function of underlying brain.

Gamma Knife sterotactic radiosurgery has been used since the 1960s for the treatment of vascular malformations and cerebral neoplasms. UVa is currently participating in a study to determine whether this will be effective for patients with temporal lobe epilepsy.

Recent advances in the treatment of epilepsy offer the possibility of cure and improved quality of life. The key to effective epilepsy management is the appropriate use of these advances. Pharmacologic management should be aimed at providing the most appropriate AED for each specific seizure type. Patients should be considered for epilepsy surgery early in their course of treatment.

Nathan B. Fountain, M.D, nbf2p@virginia.edu

Stacey Epps, M.D.

Shorvon S, Dreifuss F, Fish D, Thomas D, eds: The Treatment of Epilepsy. Cambridge: Blackwell Science Ltd; 1996.

Foldvary N, Bingaman WE, Wyllie E. Surgical treatment of epilepsy. Neurologic Clinics. 2001 May;19(2):491-515.

Epilepsy Foundation. On-line information [www.efa.org]

Click here to go to the University of Virginia Health System's Epilepsy pages.