An official website of the United States government
NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.
Bromfield EB, Cavazos JE, Sirven JI, editors. An Introduction to Epilepsy [Internet]. West Hartford (CT): American Epilepsy Society; 2006.
An evaluation for epilepsy surgery is appropriate for anyone with seizures that may be focal in origin, that are continuing to occur despite treatment with antiepileptic drugs (AEDs), and whose quality of life is significantly impaired by epilepsy. Although the exact number of medication trials must be individualized, initial consideration is reasonable after 2 monotherapy trials with first line AEDs, and possibly one trial with duo-therapy (combination of 2 AEDs). (Slide 2)
Candidates for Epilepsy Surgery
The goal of the presurgical evaluation is to determine if the patient has a single epileptogenic focus that is not in "eloquent" cortex, and can therefore be resected without causing an unacceptable neurological deficit. The most common location of seizure onset in adults is the temporal lobe, especially the medial temporal lobe (hippocampus). This is also the seizure location most amenable to surgical cure. (Slide 3)
Presurgical Evaluation
This is cornerstone of the epilepsy surgery evaluation. Continuous EEG with synchronized video monitoring is performed in an epilepsy monitoring unit until the patient has their typical spells. Medication withdrawal is often necessary to help elicit spells more quickly.
The first step is to confirm that the patient's habitual spells are indeed epileptic. A significant proportion of patients referred for epilepsy presurgical evaluation do not have epileptic seizures (about 25%, mostly psychogenic non-epileptic seizures). Although a detailed history can be quite helpful in this regard, video/EEG monitoring of a typical spell is required to make this diagnosis definitively.
Lateralization (left vs. right hemisphere) and localization (specific region within one hemisphere) begins with standard scalp EEG recordings. Many centers utilize additional electrodes that are particularly sensitive to medial temporal discharges, including inferior temporal and anterior temporal surface electrodes and/or sphenoidal electrodes (semi-invasive electrodes inserted through the mandibular notch at the bedside).
The presence of interictal (between seizures) epileptiform discharges (EDs: sharp waves or spikes) in a single location is highly suggestive of seizure onset in that region. Computer detection of EDs is usually performed during continuous video/EEG monitoring to assist in identifying these discharges. Lack of EDs does NOT rule out epilepsy, as a significant minority of patients with epilepsy will not have them on scalp EEG (up to 10%). Medial temporal lobe epilepsy patients usually have EDs from the anterior-mid temporal lobe (electrodes F7/F8, T3/T4, as well as anterior temporal, sphenoidal, and inferior temporal electrodes if used). Bilateral temporal lobe EDs are not unusual in patients with temporal lobe epilepsy (TLE), including patients with unilateral seizures.
Ictal scalp EEG during complex partial or secondarily generalized seizures will usually show a lateralized rhythmic discharge that evolves in frequency, amplitude and location. Simple partial seizures (including auras) often have no scalp EEG correlate. The typical EEG correlate of a medial temporal lobe seizure is a rhythmic theta or alpha discharge beginning in the anterior-mid temporal electrodes.
The clinical manifestation of a seizure (seizure semiology) is also of localizing value. For example, a typical patient with TLE may have an epigastric aura, followed by a quiet period of unresponsiveness with staring, lip-smacking (oral automatisms), picking at sheets or clothes (manual automatisms), contralateral dystonic posturing, and postictal confusion and lethargy. If from the dominant hemisphere, there is usually delayed recovery of language, often with transient aphasia on testing. A typical frontal lobe seizure will occur from sleep with no warning, may show restlessness and/or prominent bilateral limb movements (such as bicycling or asymmetric tonic posturing), and will end quickly with immediate recovery; this may recur several times in one night. Occipital lobe seizures often have a visual aura, and may progress (due to electrical spread) into a temporal lobe or frontal lobe type of seizure. Parietal lobe seizures are the least common, may have a sensory aura, and tend to mimic frontal lobe seizures.
The presence of a focal epileptogenic lesion on MRI increases the chance of surgical cure significantly. (Slide 4)
Presurgical Evaluation
In patients with temporal lobe epilepsy (TLE), special views of the medial temporal lobes with thin oblique coronal cuts often reveals hippocampal atrophy and increased signal. These findings strongly correlate with pathological evidence of hippocampal sclerosis (neuronal loss and gliosis, also called mesial temporal sclerosis, or MTS) and with seizure freedom after temporal lobectomy (approximately 80% chance). Many patients with MTS have a history of febrile seizures in childhood as well. (Slide 5)
Presurgical Evaluation
MRI may also show tumors (usually low grade), vascular malformations (usually arteriovenous or cavernous malformations), or cortical dysplasia. Detection of dysplasia often requires special views and careful examination for ectopic gray matter (heterotopias), cortical thickening, and blurring of the normal gray-white junction.
Positron emission tomography (PET) utilizes an injection of radio-labeled glucose (18FDG) to measure brain metabolism. Interictal PET usually shows hypometabolism in the seizure focus, especially in TLE. Ictal PET is not practical due to the extemely short half life of the radiotracers used. PET is most useful in MRI-negative TLE, though it may be helpful in extra-temporal epilepsy as well. (Slide 6)
Presurgical Evaluation
Single photon emission computed tomography (SPECT) utilizes injection of a radio-labeled tracer of blood flow that binds on first-pass through the brain; thus, it is a snapshot of circulation at the time of injection. The tracer is stable for several hours, allowing delayed imaging. The most useful study for presurgical evaluation is an ictal SPECT, in which the injection is performed (as early as possible) during a seizure and the patient is scanned within a few hours. Ictal studies usually reveal increased blood flow at the site of seizure onset. Interictal studies often show relative hypoperfusion at the site of seizure onset. Comparing ictal and interictal studies, including quantitative subtraction, can add additional information, as can co-registration with MRI SIScom. (Slide 7)
Presurgical Evaluation
MEG is a relatively new diagnostic technique, primarily used in research settings. MEG is similar to EEG, but it detects magnetic rather than electric signals from the brain. MEG can sometimes detect epiletpiform discharges in patients with normal scalp EEGs and can be considered as a test complementary to EEG.
fMRI can be used to map motor, sensory and language functions noninvasively, and is most commonly used as part of surgical planning. fMRI can detect focal changes in blood flow and oxygenation levels that occur when an area of the brain is activated. Possible future applications of fMRI include localizing epileptiform discharges or seizures, and lateralizing memory function.
Formal neuropsychological testing is important as a pre-operative baseline, as a predictor of possible cognitive loss with surgery, and as an additional aid for localization. For example, patients with temporal lobe epilepsy tend to have memory deficits. Those with dominant TLE (usually left sided) have more prominent deficits in verbal memory compared with visual memory. Patients with average or above average memory function prior to temporal lobectomy have a higher risk of memory decline, especially with left (dominant) temporal lobectomy. (Slide 8)
Presurgical Evaluation
Named after Dr. Juhn Wada, this test helps determine the risk of postoperative memory and language deficits, and aids in seizure localization. Amobarbital is a short-acting barbiturate that is injected into the internal carotid artery (via femoral artery puncture), resulting in unilateral hemispheric anesthesia for approximately 10 minutes. During this time, memory items are given to the patient and language is tested. After recovery, recall of the memory items is tested, and then the test is repeated on the other side. A patient with unilateral TLE will usually have a significant memory asymmetry with this test, as the epileptogenic hippocampus is already dysfunctional. If the hippocampus to be resected is functioning normally on this test, the chance of a postoperative memory deficit is greater, especially in the dominant hemisphere. Most centers require demonstration of intact function of the contralateral hippocampus on this test prior to offering temporal lobectomy in order to prevent a severe postoperative amnestic syndrome. Patients with extratemporal epilepsy usually have intact memory function bilaterally.
If the seizure focus cannot be adequately localized and safely resected based on the above studies, recording seizures with intracranial EEG may be necessary. Intracranial electrodes are inserted neurosurgically, and include subdural (or epidural, though less common) strips or grids of electrodes, or parenchymal "depth" electrodes. Depth electrodes are thin probes with multiple electrodes along their length that are most commonly used to record from the hippocampus. The exact location and types of electrodes are tailored to each individual case. Intracranial electrodes, especially subdural grids, can also be used for identification of cortical areas that are important for language, movement, and sensation via cortical stimulation and/or recording of evoked potentials (see brain mapping section). (Slide 9)
Presurgical Evaluation
Epilepsy surgery can be divided based on the goals of the operation into palliative and curative procedures. Examples of curative procedures include lesional resection, lobectomy, corticectomy, and some cases of hemispheric surgery and multiple subpial transections. There is preliminary evidence that gamma knife radiosurgery may have a role in the treatment of temporal lobe epilepsy and for cavernous malformations associated with epilepsy. The primary goal of a curative surgery is for the patient to be able to lead a normal life, preferably off of all antiepileptic medications. There is gathering evidence that early surgical intervention is favorable for a variety of reasons. Becoming seizure free at a younger age may lessen the cognitive, behavioral, and psychosocial problems experienced by epilepsy patients, potentially improving societal integration. Additionally, because continued seizures may result in progressive neurologic damage over time, surgery has the potential to be neuroprotective in comparison to continued medically refractory seizures. If a hemispheric procedure is required or if "eloquent" areas of brain are within the epileptogenic zone, the potential for recovery of language and sensorimotor functions is better when patients are younger. (Slide 10)
Types of Surgical Procedures
By definition, palliative procedures only very rarely result in cessation of seizures. These surgeries may prevent the occurrence of a particularly morbid type of seizure such as drop attacks or lessen the frequency or severity of seizures. Palliation may be a desirable result in patients with seizure related injuries or with a predominance of one seizure type that can be eliminated with surgery. Examples of palliative surgery include some cases of hemispheric surgery, multiple subpial transections, disconnection procedures including corpus callosotomy. There is a continuum between likely curative and likely palliative procedure (Slide 11), and patient and family expectations must be adjusted accordingly. For example, the procedure of choice for a patient with invasive monitoring-documented mesial temporal lobe epilepsy (TLE) who has no magnetic resonance imaging evidence of mesial temporal sclerosis (MTS) is resection of the anteromesial temporal lobe through one of a variety of surgical methods (see below). The likelihood of seizure freedom in these patients is approximately 60%, in contrast to 80–90% seizure freedom if MTS is present on MRI.
Surgical Treatment of Epilepsy
The goal of epilepsy surgery is either to define and resect an area of epileptogenesis (seizure focus) or disrupt spread of seizure activity to reduce the likelihood of seizures or prevent certain seizure types. Most surgical candidates suffer from partial seizures, and many have epilepsy secondary to definable structural abnormalities. The location and nature of these lesions dictates the type of surgery that will be performed and the expected outcome. The concept of surgically remediable syndromes is important when considering people for surgical evaluation. These are medical syndromes that we have come to understand respond poorly to medical therapy and well to surgical treatment. Patients with these problems may be considered earlier for surgical intervention than some other cases of partial epilepsy. Medial temporal lobe epilepsy is the most common of these syndromes. It classically consists of a history of a complex or atypical febrile seizure in early childhood, onset of recurrent seizures in late childhood or adolescence, complex partial seizures, and evidence of hippocampal sclerosis on MRI. Lesional epilepsy, caused by lesions such as vascular malformations and cortical dysplasia, also responds well to surgical therapy in most cases.
These invasive procedures are used when the noninvasive presurgical evaluation (detailed in the previous section) is inadequate to define the epileptogenic zone reliably enough for surgery.
Depth electrodes are used primarily in cases of suspected medial temporal lobe epilepsy. They are thin cables with cylindrical contacts lying along their distal ends. They are placed within the brain parenchyma, usually in the hippocampus and amygdala. These electrodes are used when medial temporal lobe epilepsy is suspected but cannot be reliably lateralized due to rapid propagation from one medial temporal lobe to the other and to distinguish medial from lateral (neocortical) temporal lobe seizure onset. They are usually placed stereotactically through burr holes along the long axis of the hippocampus (entering from the occipital lobe), or orthogonal to the long axis of the hippocampus (entering through the lateral temporal lobe). They may be used in combination with subdural strip electrodes to record from the lateral temporal cortex as well.
Subdural electrodes are used to record from the surface of the brain. They are most commonly used to delineate a region of seizure onset in neocortical epilepsy, although subdural electrodes that record from the parahippocampal gyrus may be used in suspected medial temporal lobe epilepsy. The electrodes are thin discs of platinum or stainless steel embedded in a thin sheet of plastic. Unless they are configured in thin strips, they require a craniotomy for placement. These electrodes can be used both to record the area of seizure onset and to perform extraoperative stimulation mapping of cortical function in the brain tissue beneath the electrodes.
Anterior temporal lobectomy (ATL) is used to treat medial temporal lobe epilepsy. Medial temporal lobe epilepsy is the most common form of intractable epilepsy in adolescents and adults. Seizures usually arise in the hippocampus, although they may arise from the amygdala or parahippocampal gyrus. MTS is the pathological hallmark of medial temporal lobe epilepsy. It is characterized by loss of hippocampal neurons in a particular pattern (area CA1 is most severely involved), gliosis, and synaptic reorganization in the inner molecular layer of the dentate gyrus (mossy fiber sprouting). MTS can frequently be detected non-invasively using magnetic resonance imaging. In the case of typical medial temporal lobe epilepsy, a standardized anterior temporal lobectomy is commonly performed. As described by Spencer and colleagues (1984), this resection includes a small amount of the anterolateral temporal lobe, the majority of the amygdala, the uncus, and the hippocampus and parahippocampal gyrus back to the level of the collicular plate. An alternate approach is to define the limits of the resection using physiologic criteria for each individual patient. This requires implantation of chronic subdural electrodes if ictal onset is desired (most useful) or intraoperative electrocorticography (ECOG) if interictal abnormalities are used (less reliable). This approach is used if there are concerns that the seizure focus may extend beyond the medial temporal lobe.
Side effects of ATL include visual field defects in the contralateral superior quadrant, due to damage to the fibers of Meyer's loop, and memory deficits. Functionally significant visual field deficits are uncommon using the various techniques that spare all but the anterior 3–4 cm. of the lateral temporal lobe. Measurable worsening of verbal memory can result after language-dominant ATL. In most cases, this risk is determined by preoperative functioning. If verbal memory is intact preoperatively based on neuropsychological testing and the Wada test, a more pronounced decrement can be expected after resection of the dominant medial temporal lobe. If preoperative verbal memory is impaired, then little or no decrement is seen. In most cases, because MTS is associated with preoperative verbal memory deficits, most patients who undergo ATL show little or no significant deterioration of memory. The utility of temporal lobe surgery for intractable epilepsy compared to continued treatment with antiepileptic drugs was recently demonstrated in a prospective, randomized, controlled trial (Wiebe et al, 2001). Eighty patients were randomized to surgery or medical treatment for one year. At one year, those undergoing surgery had a much higher rate of seizure freedom (58% versus 8%) and a significantly better quality of life.
Lesionectomy refers to surgical resections aimed at curing epilepsy by removing strucural brain lesions — most commonly malformations of cortical development, low-grade neoplasms, or vascular malformations. The surgical approach depends on the location of the lesion. Intraoperative frameless stereotaxy has been a major technological improvement in intraoperative localization of subtle cortical lesions and in correlating the location of lesions with physiologic data acquired through subdural electrodes (Slide 12). Magnetic resonance images used in frameless stereotactic localization of an area of focal cortical dysplasia at the base of the central sulcus (center of cross-hairs) enabled adequate coverage with subdural electrodes of a lesion that was not visible from the surface of the brain. There is ongoing debate over when the resection should be limited to the lesion itself and when additional adjacent cortex should be removed in association with the lesion. The specific pathology involved, eloquence of the adjacent cortex, and duration of the epilepsy determine the best approach for each patient. Cavernous malformations are examples of well-circumscribed lesions where the region of epileptogenesis is commonly limited to the immediately adjacent cortex. In contrast, cortical dysplasia stands out as a pathological entity that may pose problems in determining resection boundaries, and a combination of structural, metabolic, and physiologic data may need to be employed in these cases.
Surgical Treatment of Epilepsy
Resection of cortex outside the medial temporal lobe is referred to as a topectomy, a corticectomy or a neocortical resection. The boundaries of these resections are typically determined by recording the area of seizure onset with chronically implanted subdural electrodes. Again, the surgical approach depends on the location of the focus. In the absence of pathological abnormalities, extratemporal resections represent the poorest outcome group of the surgical resections. Because suspected regions of epileptogenesis may involve eloquent cortex in these cases, mapping of cortical function is often an important part of the diagnostic work-up. This may include extra-operative techniques such as functional MRI, magnetoencephalography, and cortical stimulation and somatosensory evoked potential (SSEP) mapping through subdural electrodes as well as intra-operative cortical stimulation and SSEPs.
Hemispherectomy is used in patients where seizures arise over most or all of one cerebral hemisphere. Processes that result in this condition usually also produce severe damage to the involved hemisphere early in development. These include diffuse cortical dysplasia, pan-hemispheric Sturge-Weber syndrome, large perinatal infarcts, hemimegalencephaly, and Rasmussen's encephalitis. The goal of hemispherectomy is to remove or disconnect all of the cortex of one hemisphere from the rest of the brain. Anatomical hemi-spherectomy is used to resect the hemispheric cortex in its entirety, whereas a functional hemispherectomy (Tinuper et al., 1988) removes the temporal lobe and central cortex but preserves some of the frontal and occipital cortex (Slide 13). The upper image shows a hemisphere that had been severely injured by an infection early in childhood. The lower image shows the extent of the cortical resections in the temporal lobe and central cortex. Disconnection of the residual frontal and occipital cortex is accomplished by transecting the fibers of the white matter while working within the lateral ventricle (not shown). The white matter connections to the residual cortex are completely divided so that, even though it is still viable, it has no functional effect on the rest of the brain. More recent modifications of the technique involve removing only enough cortex to gain access to the lateral ventricle and performing a complete disconnection of the hemisphere working from within the ventricle.
Surgical Treatment of Epilepsy
Although it may sound surprising, new deficits after hemispherectomy are quite rare. This is because most of the conditions that are treated with this surgery result in a devastating injury to the hemisphere early in development. The person's brain then develops by relying on the "good" hemisphere. Following major childhood injuries, the pediatric brain can often achieve substantial recovery of function due to the increased plasticity of the young brain. A notable exception to this scenario is Rasmussen's encephalitis in an older patient, where a progressive and relentless loss of function of the hemisphere may be seen. Even in early injuries there are limits to transfer of function to the contralateral hemisphere, and loss of fine motor control in the contralateral hand and some degree of contralateral visual field defect are to be expected.
The multiple subpial transections (MST) procedure was developed to treat epilepsy arising from cortex that cannot be resected (Morrell et al., 1989). The underlying rationale is that disruption of horizontal connections that run within the cortex which are vital for synchronizing neural activity (a key feature of seizures) can be accomplished without affecting the ascending and descending fibers that are critical for the normal functioning of the cortex. The technique involves the use of a small hook that cuts through the gray matter while leaving the overlying pia and surface blood vessels intact. These transections are made at right angles to the long axis of the gyrus at 5 mm intervals. Since all of these patients have extratemporal epilepsy, most centers determine the boundaries of the area to be transected from ictal onsets recorded with chronically implanted subdural electrodes. Centers vary on whether they use intraoperative ECoG and cessation of interictal spiking in the area as a physiological end-point of the procedure. Several reports have documented that MST can be performed in primary motor, sensory and language cortices without producing a significant permanent new deficit. Responses in terms of seizure control vary widely between reported series with 20–70% seizure free (Spencer et al, 2002). Efficacy for treatment of Landau-Kleffner Syndrome (acquired epileptic aphesia) has been very promising in a small number of patients (Morrell et al, 1995), but is difficult to assess fully as MST's are often combined with adjacent cortical resections.
Transection of the corpus callosum is intended to disrupt the rapid spread of certain seizures from one hemisphere to the other. Although indications for the procedure vary between centers, callosotomy is most commonly performed for atonic drop attacks (Spencer et al., 1988). These seizures are extremely injurious because of their rapid onset and often produce multiple injuries over time due to unprotected falls. They are most commonly seen in the setting of Lennox-Gastaut Syndrome, a severe symptomatic generalized epilepsy syndrome with multiple seizure types. Because of increased side effects with complete callosotomy, the initial surgery usually sections the anterior 66–75% of the callosum. If seizure response is less than expected or transient, a second procedure is sometimes performed to complete the callosal section. Complete callosotomy at the initial surgery is usually reserved for patients whose baseline neurologic function is so impaired that a major disconnection syndrome would not affect their quality of life. Complications of callosal sectioning can be divided into early or transient and permanent effects. The early or transient disconnection syndrome can include mutism and inability to initiate movement of one or more limbs. This can be seen after partial or complete callosotomy, lasts one to several weeks, and then resolves. The more severe, complete disconnection syndrome is rare after partial callosotomy but more common after complete sectioning. It includes complete inability to transfer sensory information from one hemisphere to the other and major problems with coordination and motor control of the non-dominant limbs. In some cases the non-dominant arm may even act in an autonomous fashion and antagonize the actions of the dominant arm. Seizure response is best for atonic drop attacks although some centers use callosotomy for a wider range of seizure types. Some series have shown an increase in partial seizures after callosotomy, presumably due to a loss of inhibitory input to a cortical focus from the contralateral hemisphere via the callosum.
Surgical therapy offers hope for improved quality of life for a wide variety of patients with intractable epilepsy. Expected outcome from surgery is dependent on the type of surgery performed and the location and nature of the pathological substrate. Advances in structural, functional and metabolic imaging are greatly enhancing our ability to define the extent and nature of epileptogenic lesions prior to surgery and, therefore, allowing us to make better decisions in surgical candidacy and operative planning.
Antiepileptic drugs (AEDs) are the primary treatment for epileptic seizures and 60–80% of newly diagnosed patients will achieve seizure control with AED therapy. The sizable group of remaining patients are defined as having medically refractory or intractable epilepsy. These patients are not evenly distributed within the epilepsy population. Patients with refractory partial seizures are quantitatively the largest group, representing 72% of those with chronic epilepsy in a recent population survey from the United Kingdom (Hart et al, 1995). Young children with catastrophic epilepsies such as the Lennox-Gastaut syndrome are also disproportionately represented. The vagus nerve stimulator (VNS) was developed as a treatment for medically refractory epileptic seizures. (Slide 14)
Vagus Nerve Stimulation
The device is a standard pacemaker consisting of a generator (housing a lithium battery and electronics), and a lead wire. The generator is implanted in a subclavicular pocket. The lead wire is tunneled into the left carotid sheath via a transverse or longitudinal neck incision, and the spiral endings of the leads are attached to the left vagus nerve. The left vagus is used due to a lower percentage of efferent fibers to the atrioventricular node. The device is programmed with a telemetry `wand', held over the skin overlying the generator and connected to a portable computer or handheld PDA. Settings for current, frequency, duty cycle, etc. are selected in software. A typical cycle includes 30 seconds of 30 Hz stimulation followed by a 3–5 minute off period, 24 hours/day. The device can also be manually activated by the patient or caregiver, using a pocket magnet.
Two multicenter controlled clinical trials (The VNS Study Group, 1995 and Handforth et al 1998) lead to FDA approval of VNS in 1997 as adjunctive therapy for the treatment of partial seizures in patients 12 years of age and older. The design of these trials was similar to those used in recent AED trials, and the results are shown in Slide 15. Each bar represents the mean percent change in seizure frequency for the final 12 weeks of VNS as compared to a 12 week pretreatment baseline. Differences between the `high' and `low' stimulation groups were statistically significant. In both studies, high level stimulation produced a significantly greater reduction in seizures than did low level stimulation. Slide 16 shows responder rates during the 12 weeks of active stimulation. Each pair of bars represents the percentage of patients who achieved at least a 50%, 75%, or 100% decrease in seizure frequency during the final 12 weeks of VNS as compared to the 12 week pretreatment baseline. Only one patient became completely seizure-free. VNS is therefore considered a palliative therapy and is not curative. Slide 17 compares responder rates for several of the newer AEDs with VNS. Odds ratios (OR) and associated 95% confidence intervals for at least a 50% decrease in seizure frequency in response to gabapentin, lamotrigine, topiramate, tiagabine, or VNS. All pooled data are from controlled, randomized trials in patients with medically refractory partial seizures. AED data are from Marson et al (1997).
Vagus Nerve Stimulation Percentage Change – All Seizures
Vagus Nerve Stimulation "Responder Rates"
Metaanalysis of AEDs and VNS Efficacy
There are limited data from open-label trials only on the treatment of medically intractable generalized epilepsy and for the treatment of epilepsy in children. Promising preliminary results have been reported in juvenile myoclonic epilepsy, absence epilepsy, and Lennox-Gastaut Syndrome.
Most patients experience some hoarseness when the stimulator is firing (due to recurrent laryngeal nerve activation). This is usually mild, current dependent, and decreases with time. No cardiac, respiratory or gastrointestinal effects were seen during the controlled trials. Wound infections were encountered in approximately 1% of cases. Investigation of the syndrome of sudden unexpected death in epilepsy (SUDEP) revealed that this syndrome is no more frequent in patients receiving VNS than would be expected in a similar population of patients with medically refractory epilepsy (Annegers et al, 1998).
VNS may provide an alternative for patients whose seizures have failed to respond to AED therapy (or who are intolerant of AEDs) and who are not optimal candidates for curative epilepsy surgery. Just as AEDs are often used to treat disorders other than epilepsy (mania, headache, neuropathic pain), VNS may also have such a role. Despite encouraging preliminary observations, clinical trials of VNS for depression have not yet produced convincing evidence of efficacy (Schachter 2004).
Head injury is a major cause of neurologic morbidity and mortality throughout the world. It is one of the main causes of "remote symptomatic epilepsy," possibly causing 20% of such epilepsy and 5% of all epilepsy. (Slide 18)
Epilepsy and Head Injury
The risk of seizures and later epilepsy is strongly related to the severity of the head injury. About 2% of patients with injuries severe enough to cause loss of consciousness, and 7–15% of those injured seriously enough to be admitted to a hospital, go on to develop later epilepsy. Series with higher risks are usually those in which patients were selected by hospitalization for more severe trauma. Early seizures (those in the first week after injury) may occur more often in children, but they appear to be associated with later epilepsy more often in adults. Only a minority of patients with early seizures go on to develop epilepsy.
Among the primary risk factors for seizures and epilepsy are early seizures, penetrating head injuries, parenchymal hemorrhages or subdural hematoma, and a low Glasgow coma scale score (3–8). Risk factors found at least occasionally were cortical injuries, larger volume of tissue lost, depressed skull fractures, retained metal fragments, and longer duration of loss of consciousness. Among the highest reported risks were in military series from Vietnam (Salazar et al, 1985), but they are subject to substantial selection bias because patients with more severe disabilities and epilepsy are more likely to require late treatment and be ascertained during prolonged follow up than are people without seizures. Penetrating head injuries led to epilepsy in up to 53% of these patients. Persistent focal neurologic deficits were particularly ominous: 90% of patients with residual aphasia had epilepsy. In general, the EEG has not been particularly informative in predicting risk, but in one series, development of an epileptiform focus one month after injury predicted later epilepsy in 50% of cases (Angeleri et al, 1999). (Slide 19)
Epilepsy and Head Injury
In a Mayo Clinic study by Annegers and colleagues, patients with severe injuries (with contusion, intracranial hematomas, focal neurologic deficits, or more than 24 hours of unconsciousness or post traumatic amnesia) had an 11.5% chance of developing epilepsy within 5 years. Those with moderate injury (skull fractures or at least 30 minutes of unconsciousness or amnesia) had a 1.6% risk (still increased from the general population) and those with milder injuries had no elevated risk. Those with severe injury and early seizures had a 36% chance of later epilepsy. (Slide 20)
Epilepsy and Head Injury
About 80% of patients who develop late seizures do so within 2 years, 60% in the first year, but 12% of all epilepsy due to head injury present more than 10 years after the injury, and the risk extends to at least 20 years (Annegers et al, 1998).
Seizures just after head injury can be considered provoked or symptomatic, i.e. the consequence of a known cerebral insult. These acute symptomatic seizures are expected to cease with resolution of the precipitating cause or illness. Remote symptomatic epilepsy, i.e., recurrent unprovoked seizures due to a specific brain injury but occurring long after it, however, would not be expected to resolve and usually requires treatment.
Whether to use antiepileptic drugs (AEDs) in the setting of head injury depends on the anticipated likelihood of seizures and their harmful consequences balanced by the likely effects of AEDs. Early seizures are typically treated with short-term AEDs (assuming that the correct diagnosis of an epileptic seizure has been reached), but long-term treatment assumes a likely development of recurrent seizures (epilepsy).
Prophylactic AED use following head injury has been very common practice but is quite controversial. Early studies of AEDs after head injury were impressively favorable but almost always uncontrolled. In one of the best, more recent, and well controlled studies, Temkin and colleagues randomized and treated over 400 patients within 24 hours of injury with either phenytoin (adjusted to therapeutic levels) or placebo. The injuries included contusion or hematoma on CT scan, depressed skull fracture, penetrating wound, early seizure before the 24 hour time period, or GCS score < 10. Thus, patients were selected for having a significant risk of later seizures. Seizures occurred in 14% of those on placebo and 4% on phenytoin within the first week, establishing efficacy, at least in serious head injury. Nevertheless, there was no significant difference in later seizures or overall outcome at one or two years, arguing that long term treatment was of no benefit. The inadvisability of prolonged treatment derives from its lack of efficacy in preventing late seizures and also from possible cognitive and other side effects, although most of these were relatively minor and appeared to clear within days of medication cessation. Most experts recommend AED prophylaxis for 1–2 weeks only. (Slide 21)
Head Injury and Prophylactic AEDs
Once a single late posttraumatic seizure has occurred, the treatment decision becomes easier. In a continuation of the prophylaxis trial above, 86% who had a single seizure after the first week went on to have at least one more seizure, 50% within the first month. This was despite AED treatment in most, thus likely underestimating the risk. Even those with no clear risk factors had recurrence in 65% of cases.
Early on, it was speculated that prophylactic or early treatment of epileptic seizures might retard the development of later recurrent, unprovoked seizures, i.e. epilepsy. This would be truly prophylactic treatment rather than just seizure suppression. Well controlled series have found no evidence of any such benefit, so true prophylaxis does not yet appear possible (Temkin, 2001).
In the end, AED use is appropriate even before a seizure if the injury is severe, e.g. for patients with coma, penetrating injury, or hemorrhage. Treatment suppresses seizures, at least during the first week. Afterward, there does not appear to be any efficacy or reason for continuation. Subsequently, even a single unprovoked seizure predicts epilepsy in most patients and warrants treatment with AEDs. The duration of treatment, however, remains uncertain. A seizure that develops years after a head injury or after a mild head injury may be unrelated to the head injury. Treatment decisions depend upon a thorough examination of the clinical situation surrounding the seizure.
All content of An Introduction to Epilepsy, except where otherwise noted, is licensed under a Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Your browsing activity is empty.
Activity recording is turned off.
See more...
