Resective and ablative surgical treatment of epilepsy in adults

INTRODUCTION AND ROLE OF SURGICAL THERAPY — 

Epilepsy is one of the most common chronic neurologic disorders, and approximately 20 to 30 percent of patients with epilepsy will have medically and socially disabling seizure disorders. Such patients are at increased risk for serious morbidity and mortality, including cognitive disorders, depression, physical trauma, and sudden unexpected death in epilepsy.

Surgical therapy is an important and underutilized treatment in patients with drug-resistant focal epilepsy.

Surgical procedures for epilepsy range from focal resection of the epileptogenic cortex to minimally invasive techniques (laser interstitial thermal therapy, stereotactic radiosurgery). Interventions that remove or isolate the cortex of a grossly diseased hemisphere (functional hemispherectomy, anterior corpus callosotomy, multiple subpial transections) are most often performed in children and are not discussed further here.

The goals of treatment for individuals with drug-resistant epilepsy are to render the patient seizure-free, avoid treatment-related adverse effects, and allow the individual to become a participating and productive member of society. In general, only complete resection of the epileptogenic brain region offers the chance of long-term seizure freedom.

This topic will discuss the surgical treatment of drug-resistant focal epilepsy in adults utilizing focal cortical ablation. The presurgical evaluation for epilepsy surgery is reviewed separately. (See "Epilepsy surgery: Presurgical evaluation".)

Other aspects of epilepsy management in adults and epilepsy surgery in children are discussed elsewhere:

●(See "Overview of the management of epilepsy in adults".)

●(See "Evaluation and management of drug-resistant epilepsy".)

●(See "Seizures and epilepsy in children: Refractory seizures", section on 'Epilepsy surgery'.)

SURGICALLY REMEDIABLE EPILEPSY SYNDROMES — 

In well-selected patients with drug-resistant epilepsy (DRE) (eg, those with focal epilepsy), epilepsy surgery is superior to medical therapy [1,2]. All patients with DRE who are potential surgical candidates should be evaluated by a comprehensive epilepsy center. The goals of the evaluation are to exclude seizure mimics and confirm the diagnosis of epilepsy, identify surgical candidates, define the epileptogenic zone and eloquent cortex, and look for factors that may preclude surgery. (See "Epilepsy surgery: Presurgical evaluation".)

Mesial temporal lobe epilepsy — Mesial temporal lobe epilepsy is the most frequently encountered surgically remediable epilepsy syndrome in adults. Patients typically experience focal seizures, also known as complex partial seizures, with or without aura or tonic-clonic seizures.

The most common pathologic substrate associated with temporal lobe epilepsy is hippocampal sclerosis (mesial temporal sclerosis), which is characterized by selected focal neuronal loss and gliosis in the hippocampus, predominantly affecting CA1, CA3, and the dentate granule cell layer. Hippocampal sclerosis is the most commonly encountered histopathologic diagnosis at the time of epilepsy surgery in adults and accounts for approximately 45 percent of all resections [3]. (See "Focal epilepsy: Causes and clinical features", section on 'Seizure semiology'.)

Patients with mesial temporal lobe epilepsy due to tumors or other structural lesions in the temporal lobe are discussed below in the context of lesional epilepsy surgery. (See 'Lesional epilepsy' below.)

Those with extrahippocampal temporal lobe epilepsy and a normal brain magnetic resonance imaging (MRI) are discussed in the context of neocortical epilepsy with normal brain MRI. (See 'Focal epilepsy with normal brain MRI' below.)

●Anterior temporal lobe resection – The most common surgical procedure for mesial temporal epilepsy is resection of the anterior temporal pole, hippocampus, and part of the amygdala. The posterior extent of the anterior temporal resection is measured to minimize risk to the visual radiations and language cortex (4.0 to 4.5 cm back from the temporal pole on the dominant side, and 5.0 to 5.5 cm on the nondominant side). Cortical resection may also be tailored based on language mapping and intraoperative electrocorticography [4], as language areas in the temporal cortex are variable.

●Selective amygdalohippocampectomy – Selective amygdalohippocampectomy has been explored as an alternative to anterior temporal lobectomy that spares the temporal lobe neocortex. Seizure control rates might be similar with a selective approach, but potential differences in neurocognitive outcomes have not been well studied [5]. There are conflicting results regarding the effectiveness of the specific operative strategies to achieve seizure freedom in individuals with mesial temporal lobe epilepsy. A 2018 meta-analysis noted that the only direct evidence comparing surgical efficacy of selective amygdalohippocampectomy with anterior temporal lobectomy comes from observational studies and is therefore of low quality [5]. Importantly, most individuals experience a significant reduction in seizure frequency with either operative approach.

●Radiofrequency ablation (RFA) and laser interstitial thermal therapy (LITT) – Minimally invasive techniques are being increasingly utilized in the management of drug-resistant focal epilepsy [6-10]. In a systematic review of radiosurgery for drug-resistant focal epilepsy, which included 16 studies and 170 patients, the proportion of patients who were seizure-free or experienced rare seizures after treatment (ie, Engel Class I to II) was 58 percent [6]. However, subsequent surgery related to complications or recurrent seizures was required in 20 percent. A systematic review and meta-analysis of surgical strategies for drug-resistant mesial temporal lobe epilepsy found that selective amygdalohippocampectomy and anterior temporal lobectomy were associated with better seizure outcomes compared with RFA or LITT [10]. The proportion of patients who were seizure-free or near seizure-free (ie, Engel Class I) was 66 percent with selective amygdalohippocampectomy and 69 percent with anterior temporal lobectomy compared with 57 percent with LITT and 44 percent with RFA. However, individuals undergoing the minimally invasive techniques may have fewer major complications, especially with naming or verbal memory deficits [10].

●Efficacy in the first year or two – Patients with temporal lobe epilepsy and localization of the epileptogenic zone to the amygdala and hippocampus are extremely good candidates for resective surgery, which is associated with superior outcomes compared with continued medical management.

This was demonstrated by a randomized, controlled trial involving 80 patients with temporal lobe epilepsy whose seizures were poorly controlled with medical therapy [11]. Approximately three-quarters of the patients had MRI findings consistent with hippocampal sclerosis, 10 to 15 percent had structural lesions such as tumors or vascular malformations, and approximately 15 percent had normal MRI scans. Patients were randomly assigned to surgery or continued antiseizure medication treatment. At one year, the cumulative proportion of patients who were free of seizures that impaired awareness was significantly greater in the surgery group than the medical group (58 versus 8 percent). Quality-of-life ratings were also higher in the postsurgical group.

A second randomized, controlled study was stopped early because of slow enrollment [12,13]. After two years of follow-up, seizure remission occurred in 11 of 15 patients assigned to surgical treatment and none of 23 patients assigned to medical management.

Based on these randomized trials and data from observational studies, the most important preoperative predictors of seizure freedom after anterior temporal lobe resection for mesial temporal lobe epilepsy include [14-21]:

•Presence of a focal brain lesion on MRI

•Presence of unilateral mesial temporal sclerosis in the temporal lobe of seizure origin

•Presence of a localized temporal lobe positron emission tomography (PET) abnormality, even if brain MRI is normal

•Electroencephalography (EEG) data showing concordant location of ictal onset and interictal epileptiform discharges

•Shorter preoperative seizure duration

●Factors associated with long-term seizure control – Postoperatively, the strongest predictor of long-term seizure control is freedom from seizures in the first year after surgery, particularly generalized seizures or focal seizures with altered or impaired awareness [15,22-26]. The presence of interictal epileptiform discharges on an EEG performed within the first few years after surgery has been associated with an approximately threefold higher risk of recurrent seizures [27,28].

Taken together, patients with MRI scans showing hippocampal sclerosis who undergo anterior temporal lobectomy for intractable epilepsy have the most favorable seizure outcomes, with approximately 65 to 75 percent of patients remaining continuously seizure-free or having only auras up to 10 years after surgery [29-31]. An additional 10 to 15 percent of patients will have seizures postoperatively but eventually achieve terminal remission. In highly selected series limited to patients most likely to do well with surgery (ie, positive MRI findings, pathologically proven hippocampal sclerosis, concordant EEG findings, no dual pathology or discordant preoperative data), long-term rates of seizure freedom as high as 90 percent have been reported [32].

Focal epilepsy with normal brain MRI — The surgical management of focal seizures of neocortical origin (ie, extrahippocampal) can be challenging because of difficulty defining the boundaries of the epileptogenic zone that must be resected for seizure freedom. There are also increased concerns regarding clinically functional cortex, including motor, vision, and language areas of the brain.

●Clinical manifestations – The clinical manifestations of neocortical epilepsy depend on the area of cortex involved. Seizures arising from functional cortex can be localized based on neurologic symptoms that occur at seizure onset or during the postictal state. Compared with mesial temporal lobe seizures, lateral temporal neocortical seizures are more likely to manifest as experiential auras and less likely to manifest as epigastric auras and contralateral dystonia. Frontal lobe seizures tend to be shorter and more frequent than temporal lobe seizures, with manifestations varying from motionless staring to violent automatisms. Frontal lobe seizures are often confined to sleep. Parietal and occipital seizures typically have complex sensory symptoms such as visual hallucinations of objects or scenes.

Despite these general principles, extrahippocampal focal seizures often have varied clinical semiology; these are described separately. (See "Focal epilepsy: Causes and clinical features", section on 'Neocortical epilepsy'.)

●Seizure localization – Extrahippocampal focal seizures can be difficult to localize with scalp-recorded ictal EEG studies. In addition, ictal behaviors may relate to seizure propagation and provide few clues regarding the site of actual seizure onset. Another challenge to clinical localization in neocortical epilepsy is that seizures may be tonic-clonic without a clinically recognized focal seizure, or the focal seizure may be very brief or subtle, such as a brief stare with arrest of activity or hypermotor activity.

To adequately localize seizures and tailor resections to spare eloquent cortex, the surgical evaluation in patients with neocortical epilepsy often includes functional or metabolic imaging and long-term intracranial EEG monitoring. (See "Epilepsy surgery: Presurgical evaluation", section on 'Invasive EEG monitoring'.)

In some cases, patients who undergo a full surgical evaluation for nonlesional epilepsy, including intracranial EEG monitoring, are deemed poor candidates for surgery [33]. Common reasons include difficulty localizing seizure onset, multifocal seizures, overlap with functional cerebral cortex, or morbidity associated with chronic intracranial EEG recordings.

●Surgical procedures – Surgery for patients with neocortical nonlesional epilepsy include topectomy (ie, removal of cortex while sparing underlying white matter) and lobar and multilobar cortical resections.

●Efficacy – Even with an appropriate comprehensive presurgical evaluation, seizure outcomes tend to be less favorable in patients with focal neocortical epilepsy and normal MRI scans compared with patients who have mesial temporal lobe epilepsy due to hippocampal sclerosis or a structural lesion.

Rates of seizure freedom after such procedures range from 30 to 55 percent based on small, mostly single-center observational studies [33-38]. Predictors of better outcome include the presence of a localizing PET or single-photon emission computed tomography (SPECT) study, high concordance of the noninvasive presurgical evaluation, the presence of an aura, and complete resection of areas of ictal onset [37].

The surgical outcome of patients with a localized temporal lobe PET abnormality and a normal MRI may be equivalent to individuals with MRI-identified unilateral hippocampal sclerosis. Seventy-six percent of patients in one series with temporal lobe PET hypometabolism and a normal MRI were seizure-free following surgery [39].

For patients with mesial temporal lobe epilepsy and a normal brain MRI, rates of seizure freedom after anterior temporal lobectomy range from 50 to 60 percent. Most (70 to 87 percent) achieve at least a 75 percent reduction in seizure frequency [40-45]. In a series of 87 patients with a normal MRI undergoing anterior temporal lobectomy, 55 percent had an excellent operative outcome, defined as seizure-free or auras only [46].

Lesional epilepsy — Patients with intractable focal epilepsy due to focal brain lesions require a comprehensive epilepsy evaluation to establish the relationship between the pathologic findings and the epileptogenic zone. Common pathologic entities responsible for medically refractory lesional epilepsy include low-grade tumors, cavernous hemangiomas, and focal cortical dysplasia (FCD).

In a meta-analysis of mainly observational studies, the odds of seizure freedom after epilepsy surgery were higher in patients with lesional compared with nonlesional epilepsy (odds ratio [OR] 2.5, 95% CI 2.1-3.0) [47].

Malformations of cortical development — Malformations of cortical development (MCDs) are an important etiology for drug-resistant focal epilepsy. Among various diffuse and focal MCDs, FCD is the most common surgically remediable lesion in adults. Before widespread availability of MRI, many culprit lesions were only diagnosed at the time of post mortem examination. Even with high-resolution structural MRI, some patients with so-called MRI-negative focal epilepsy have evidence of FCD at the time of surgery.

Not uncommonly, FCD occurs in extratemporal locations and is associated with focal seizures that are difficult to localize with scalp-recorded EEG. [18F]-2-deoxyglucose PET (FDG-PET)/MRI may be particularly useful in the surgical evaluation of such patients. (See "Epilepsy surgery: Presurgical evaluation", section on 'Positron emission tomography (PET)'.)

The International League Against Epilepsy (ILAE) has developed a multi-tier histologic classification of FCD [48,49].

●FCD type I – This category is defined by architectural disorganization of the neocortex. Subtypes are characterized by abundant neuronal microcolumns (FCDIa), abnormal layering (FCDIb), and vertical and horizontal abnormalities (FCDIc). MRI studies in these patients may be normal or show very subtle blurring of the gray-white matter junction or thinning of the cerebral cortex.

●FCD type II – This category delineates dysmorphic neurons (FCDIIa) and dysmorphic neurons with balloon cells (FCDIIb). MRI studies may be normal or show blurring of the gray-white matter junction, a fluid-attenuated inversion recovery (FLAIR) signal intensity alteration, a transmantle sign, or increased cortical thickness. Abnormal cortical gyration or sulcation pattern may also be evident.

●FCD type III – This category distinguishes FCD characterized by cortical dyslamination associated with hippocampal sclerosis (FCDIIIa), tumor (FCDIIIb), vascular malformation (FCDIIIc), or a lesion acquired during early life, such as a stroke (FCDIIId).

●White matter lesions – This category delineates mild malformations of cortical development (mMCDs) that are not associated with any other lesion, such as hippocampal sclerosis, brain tumor, or vascular malformation, and primarily involve the white matter:

•mMCD with excessive heterotopic neurons

•mMCD with oligodendroglial hyperplasia in epilepsy (MOGHE)

●No definite FCD on histopathology – This category encompasses lesions with ambiguous abnormalities of cortical organization and histopathologic findings not compatible with FCD types I, II, or III.

Importantly, the diagnostic yield of MRI is dependent on the specific pathologic alterations [48]. A normal MRI study in a patient with a drug-resistant focal epilepsy does not exclude the diagnosis of FCD. Not uncommonly, MCDs associated with intractable epilepsy are extratemporal and multilobar lesions.

●Efficacy – Studies based on postsurgical histopathologic diagnosis of FCD (which included MRI-positive and MRI-negative lesions) suggested that epilepsy surgery was less effective for patients with FCD compared with surgery for patients with other lesional pathology (eg, tumors or cavernous malformations) [50]. Challenging issues in these patients include the difficulty identifying areas of FCD using MRI, the presence of extratemporal neocortical lesions, and multilobar pathology.

However, patients are generally selected for epilepsy surgery based upon presurgical evaluation that includes MRI. A systematic review and meta-analysis of patients who had epilepsy surgery for MRI-diagnosed FCD, which included 35 observational studies and 1353 patients, found that the overall rate of a favorable outcome (ie, seizure-free, or seizure-free with only auras) at ≥12 months after surgery was 70 percent (95% CI 64 to 75 percent) [51]. A favorable outcome was associated with complete lesion resection and location in the temporal lobe but was not associated with lesion extent or histologic subtype.

One center reported that 57 percent of 166 patients with FCD followed for two years or longer after surgery were seizure-free [52]. Success rates may be higher in patients with a specific form of FCD type II, in which dysplastic features are maximal at the bottom of the sulcus (referred to as a transmantle sign on MRI) [53-55].

FDG-PET coregistered with MRI may improve the surgical outcome in patients with FCD type II associated with balloon cells (Taylor-type FCD) [56]. In a study that included 23 patients who underwent stereoelectroencephalography (SEEG) and epilepsy surgery and who had pathologically verified FCD type II, MRI was negative in 13 patients and showed subtle alterations in 10 patients. FDG-PET/MRI revealed a hypometabolic zone in 22 of 23 patients. Twenty of the 23 patients (87 percent) became seizure-free following surgery.

Primary brain tumors — The incidence of seizures among patients with primary brain tumors is related to tumor type and grade and cortical localization. Low-grade, slowly growing tumors are most often associated with epilepsy. Gangliogliomas and dysembryoplastic neuroepithelial tumors (DNET) together account for approximately three-quarters of all tumors found in adults undergoing epilepsy surgery [3]. Other examples include pilocytic astrocytoma, gangliocytoma, pleomorphic xanthoastrocytoma, and oligodendroglioma.

●Imaging – Imaging features common to all these tumors include typically small size, location at or near a cortical surface, sharply defined borders, little or no surrounding edema, and, except for pilocytic astrocytoma, little or no contrast enhancement. In a series of 133 patients from one institution who underwent operations for extratemporal epilepsy, tumors were identified in 28 percent of cases. These included, in order of decreasing frequency, astrocytoma, ganglioglioma, DNET, glioneuronal hamartoma, oligodendroglioma, and oligoastrocytoma [57]. Roughly the same proportion of patients have tumors in series of patients undergoing surgery for temporal lobe epilepsy [11,58,59].

●Outcomes from surgery – Seizure outcome in patients undergoing surgical treatment for intractable focal epilepsy due to a primary brain neoplasm is typically very good. Most individuals become seizure-free or nearly seizure-free [60,61]. Although complete tumor resection is the goal, tumors associated with functional cortex (eg, perirolandic lesions) may require a subtotal excision to avoid a postoperative neurologic deficit.

Some studies have suggested that DNETs are associated with higher seizure relapse rates compared with other epileptogenic tumors [29,62]. A review of 29 relevant studies found seizure freedom rates ranging from 58 to 100 percent (median 86 percent) with a median follow-up of four years [63]. The median age at the time of surgery was 18 years. Among 12 studies that reported rates of antiseizure medication discontinuation, approximately half of seizure-free patients were off of all antiseizure medications. The most commonly identified predictors of seizure freedom were younger age at the time of surgery, shorter duration of epilepsy, and complete resection. In a separate study that included 79 patients with DNET with long-term follow-up, the rate of long-term seizure freedom was 42 percent at 10 years after surgery [29].

Vascular malformations — Cavernous malformations and arteriovenous malformations are the most common vascular lesions found in patients with focal epilepsy. Seizures are a common presenting feature of cavernous malformation. (See "Vascular malformations of the central nervous system", section on 'Cavernous malformations' and "Brain arteriovenous malformations".)

Resection typically leads to complete seizure control or significant improvement. In a case series of 168 patients with symptomatic epilepsy attributed to cavernous malformations, more than two-thirds of patients were seizure-free at three years after surgery [64]. Predictors for good outcome included mesiotemporal location, size <1.5 cm, and the absence of secondarily generalized seizures. In another study of 56 patients undergoing resection of a supratentorial cavernous malformation, 87 percent were seizure-free following surgical treatment [65].

Venous angiomas and telangiectasias are often incidental findings in patients with epilepsy and are not typically causative lesions.

Temporal encephaloceles — Temporal encephaloceles are a relatively rare cause of drug-resistant focal epilepsy but are increasingly reported in surgical series of temporal lobe epilepsy, which likely represents increased awareness and advances in neuroimaging rather than a true increase in incidence. Detection of encephaloceles is facilitated by thin-slice 3D MRI sequences and skull base computed tomography (CT) (image 1). Their presence may be related to current or prior raised intracranial pressure, based on associated imaging findings and higher prevalence in patients with obesity [66,67].

Anteroinferior/basal temporal encephaloceles, rather than posterior and lateral defects, are most commonly implicated in temporal lobe epilepsy. It remains unclear whether temporal encephaloceles represent a singular epileptogenic focus that can be removed via a lesionectomy sparing the medial structures or are a part of an extensive epileptogenic network requiring treatment with anterior temporal lobectomy and amygdalohippocampectomy.

The surgical approach to encephaloceles can vary. Extratemporal cases usually undergo lesionectomy exclusively, whereas two-thirds of temporal encephalocele cases undergo anterior temporal lobectomy and amygdalohippocampectomy [68].

SURGICAL COMPLICATIONS

Morbidity and mortality — The morbidity and mortality associated with temporal lobe resection for epilepsy treatment are low. In the Nationwide Inpatient Sample hospital discharge database, 736 patients had anterior temporal lobe resections for epilepsy performed between 1988 and 2003; overall morbidity was 10.8 percent, and there was no mortality [69]. In a multicenter surgical registry of 216 patients who had anterior temporal lobectomy and/or amygdalohippocampectomy for epilepsy between 2006 and 2014, the 30-day rate of major complications was 6.5 percent and mortality was 1.4 percent [70].

Cognitive sequelae — Epilepsy surgery poses some risk to cognitive function. Approximately one-fourth to one-third of patients develop some degree of memory loss [71,72]. Left temporal lobe resections can result in decrements in verbal memory [71,73-76], while spatial memory and learning may be affected by right-sided surgery [77].

In a six-year follow-up study of 85 patients who had serial examinations over six years, cognition continued to decline for two years following left temporal lobe resection, but it then stabilized over the next four years [78,79]. After right temporal lobe resections, cognitive scores initially improved but returned to preoperative baseline two years after surgery. Another cohort study of patients who had temporal lobe resections reported stable cognitive function for 2 to 10 years after temporal lobe resection [80]. Patients with higher presurgical abilities are at greater risk for memory decline following temporal lobectomy compared with those with lower presurgical scores [71,81-83].

In at least one comparative study, children appeared to recover lost cognitive function more quickly and more completely than adults, presumably reflecting greater neuronal plasticity [79,84]. Older age and lower baseline verbal intelligence quotient (IQ) have also been identified as risk factors in some groups [72].

The possibility of cognitive impairment after surgery needs to be balanced against the potential for control of seizures as this may be more important for functional outcomes. In one study of 138 patients undergoing epilepsy surgery, health-related quality of life (HRQOL) improved in all patients with seizure remission, regardless of neurocognitive sequelae [85]. However, in individuals without seizure remission, memory decline after surgery was associated with reduced HRQOL, which remained stable in the absence of this complication.

A preliminary, nonrandomized study in 112 patients after temporal lobectomy suggests that postoperative cognitive rehabilitation may be helpful in ameliorating some of these deficits, particularly a decline in verbal memory [86].

Visual field defects — Additional neurologic sequelae of temporal lobe surgery include visual field defects (VFDs), which are usually limited to a superior quadrant and may be detectable only by formal testing. VFDs are mainly due to injury to the inferior optic radiations (Meyer loop), which course anteriorly in the temporal lobe. In a systematic review of 76 studies that included data on over 1000 patients who underwent temporal lobe epilepsy surgery, the incidence of superior quadrantanopsia was 18 percent and the incidence of major field defects (homonymous hemianopsia) was 2 percent [87].

Risk may be higher for left-sided resections. In a prospective case series of 105 patients undergoing anterior temporal lobectomy, 16 had a new postoperative VFD; 12 of these occurred following a left-sided resection [88]. Risk may also vary according to the type of resection that is performed. In one retrospective study of 276 patients who underwent mesial temporal lobe epilepsy surgery, rates of homonymous scotoma within the central 20 degrees ranged from 56 percent after anterior temporal lobectomy to 21 percent after selective subtemporal amygdalohippocampectomy [89].

Some of the variability in rates of VFD may be related to significant interindividual variation in the anterior extent of Meyer loop, which is incompletely predicted preoperatively by current neuroimaging techniques [90-94]. Limited data suggest that use of intraoperative MRI with display of optic radiation tractography during surgery, where available, may help to minimize risk [95]. Automated imaging processing methods may improve the delineation of Meyer loop for surgical planning [96,97].

VFDs can improve somewhat from the initial postoperative defect, and driving restrictions based upon this deficit should be reevaluated up to one to two years later [98]. (See "Homonymous hemianopia", section on 'Driving'.)

Psychologic sequelae — Psychiatric problems, including depression and psychosocial adjustment difficulties, are not uncommon in patients undergoing epilepsy surgery. Risk factors include a personal or family history of psychiatric disease, low-income family and psychosocial support, and certain personality traits [99-101]. Presurgical psychiatric evaluation and psychosocial assessment and comprehensive, long-term postsurgical neuropsychologic follow-up are advised to mitigate these complications [101,102]. (See "Epilepsy surgery: Presurgical evaluation", section on 'Neuropsychologic and psychiatric assessment'.)

Other deficits — Other neurologic deficits (aphasia, cranial nerve palsy, hemiparesis) occur in 6 percent and are permanent in only half of these cases [103,104]. Death associated with epilepsy surgery is extremely rare (1 in 700) [104].

Complications in older adults — A 2024 systematic review and meta-analysis evaluated studies of epilepsy surgery included 1111 older adults (age ≥50 years) and 4111 younger adults (age <50 years) [105]. The pooled incidence of perioperative complications in older adults (reported in 22 studies) was 26.2 percent and in younger adults (reported in 11 studies) was 9.1 percent, with major complications experienced by 7.5 percent of older adults and 2.6 percent of younger adults. In studies with data for both age groups, the risk of any complication was higher for older compared with younger adults (risk ratio [RR] 2.8, 95% CI 1.5-5.4); the risk of serious complications trended higher for older adults, although the difference was not statistically significant (RR 1.3, 95% CI 0.6-2.6). The most frequent complications were language disturbance, cerebral infarctions, subdural hemorrhage, and intracerebral hemorrhage.

EPILEPSY MANAGEMENT AFTER SURGERY

Long-term seizure outcomes — Seizure control within the first year after surgery predicts more favorable longer-term seizure outcomes [22-26]. Long remissions can occur after a relapse; however, these remissions are associated with a poorer outcome than an equivalent period of complete seizure remission immediately following surgery.

Some patients will relapse after an extended period of remission, but these patients usually do not become medically intractable. Patients who develop recurrent intractability are usually identifiable within several months after surgery.

Some but not most patients are ultimately able to taper off antiseizure medications. Many other patients are successfully tapered to antiseizure medication monotherapy, which may minimize adverse effects, cost of therapy, and potential drug interactions, as discussed below.

Antiseizure medication management — The goal of surgery remains to render the patient seizure-free, with or without continued antiseizure medication treatment [106].

●Maintain antiseizure medication immediately after surgery – Most individuals are not considered for antiseizure medication withdrawal unless they are seizure-free for at least one year following surgery. Those with auras or seizure activity following surgery should probably remain on antiseizure medication therapy unless they are seizure-free for two to five years. However, limited data suggest that the incidence of recurrent seizures is not affected by the duration of postoperative antiseizure medication treatment [107].

Even if all antiseizure medications cannot be tapered, many patients are successfully tapered to antiseizure medication monotherapy, which may minimize adverse effects, cost of therapy, and potential drug interactions. In a retrospective cohort study that included 202 adults who had ablative epilepsy surgery, the rates of antiseizure medication reduction at 2 and 15 years were 47 and 46 percent, respectively, while the rates of antiseizure medication discontinuation at 2 and 15 years were 2.5 and 12.5 percent, respectively [108].

●When to consider tapering or stopping antiseizure medication – There are compelling reasons for patients to be considered for antiseizure medication taper and withdrawal following epilepsy surgery. These include antiseizure medication adverse effects, potential drug interactions, cost of therapy, and pregnancy-related considerations regarding teratogenesis. Such factors are weighed against the risk of recurrent seizures, which may affect legal driving status.

Patients, families, and caregivers understandably hope that antiseizure medications can be tapered and withdrawn after surgery. Patients often believe that medication is unnecessary if they have had "successful" surgery, and thus perceive continued use of antiseizure medications as an unwanted reminder of a disease process that is no longer an active issue for them.

There are scant prospective or randomized data evaluating discontinuation of antiseizure medications following epilepsy surgery in adults. In a review of six retrospective studies including 611 children and adults followed for one to six years postoperatively, the mean seizure recurrence rate was 34 percent after planned discontinuation of antiseizure medications in patients who were seizure-free after epilepsy surgery (most often temporal lobe surgery) [107]. The seizure recurrence rate increased during years one to three of follow-up. No factors were identified that predicted the risk of relapse after tapering antiseizure medications, despite looking at factors such as duration of epilepsy, age at onset of epilepsy, and preoperative MRI. However, this finding is limited by the retrospective nature of the available studies.

Even if all antiseizure medications cannot be tapered, many patients are successfully tapered to antiseizure medication monotherapy, which may minimize adverse effects, cost of therapy, and potential drug interactions. In a retrospective cohort study that included 202 adults who had ablative epilepsy surgery, the rates of antiseizure medication reduction at 2 and 15 years were 47 and 46 percent, respectively, while the rates of antiseizure medication discontinuation at 2 and 15 years were 2.5 and 12.5 percent, respectively [108].

●How to taper antiseizure medication – We typically obtain antiseizure medication levels and perform an EEG prior to antiseizure medication taper and withdrawal. There is controversy regarding the utility of EEG postoperatively to predict the likelihood of seizure recurrence with drug discontinuation. The emergence of epileptiform activity during the drug taper may be associated with an increased risk of seizure activity.

When the decision is made to taper, an antiseizure medication is usually withdrawn over several weeks. Long tapering schedules (eg, a year or longer) have not been shown to be beneficial [109].

●Seizure recurrence – Most, but not all, individuals who have seizure activity associated with antiseizure medication termination will become seizure-free again with reinstitution of medical treatment. In a review of retrospective studies, more than 90 percent of adults who had recurrent seizures regained control with reinstitution of previous antiseizure medication therapy [107].

TRENDS IN EPILEPSY SURGERY — 

Despite a large and growing body of evidence supporting the benefit of epilepsy surgery in carefully selected patients with drug-refractory epilepsy, surgery remains underutilized [110-115]. A population-based study utilizing the United States Nationwide Inpatient Sample found that there were 6653 resective surgeries from 1990 to 2008, and there was no growth trend over this time period [110].

At the same time, there is evidence that the patient population undergoing surgical treatment at large epilepsy centers has changed. In several studies, the absolute number of anterior temporal lobe resections, as well as the percentage of epilepsy surgeries comprising anterior temporal resections, has decreased over time [116-118]. The reasons for decreased utilization of the most common operative procedure for epilepsy are not clear. At many epilepsy centers, an increasing number of patients referred for consideration of surgery have negative MRI studies, multifocal seizures, or indeterminate outpatient diagnostic evaluations [119].

SOCIETY GUIDELINE LINKS — 

Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Seizures and epilepsy in adults".)

INFORMATION FOR PATIENTS — 

UpToDate offers two types of patient education materials, "The Basics" and "Beyond the Basics." The Basics patient education pieces are written in plain language, at the 5^th to 6^th grade reading level, and they answer the four or five key questions a patient might have about a given condition. These articles are best for patients who want a general overview and who prefer short, easy-to-read materials. Beyond the Basics patient education pieces are longer, more sophisticated, and more detailed. These articles are written at the 10^th to 12^th grade reading level and are best for patients who want in-depth information and are comfortable with some medical jargon.

Here are the patient education articles that are relevant to this topic. We encourage you to print or e-mail these topics to your patients. (You can also locate patient education articles on a variety of subjects by searching on "patient info" and the keyword(s) of interest.)

●Basics topics (see "Patient education: Seizures (The Basics)" and "Patient education: Epilepsy in adults (The Basics)")

●Beyond the Basics topic (see "Patient education: Seizures in adults (Beyond the Basics)")

SUMMARY AND RECOMMENDATIONS

●Surgically remediable epilepsy syndromes – Several different epilepsy syndromes may benefit from epilepsy surgery. (See 'Surgically remediable epilepsy syndromes' above.)

•Mesial temporal lobe epilepsy – In adult patients, mesial temporal lobe epilepsy secondary to mesial temporal sclerosis is the most frequently encountered surgically remediable epilepsy syndrome. The most common surgical procedure is an anterior temporal resection in which the temporal pole is removed along with the hippocampus and part of the amygdala. The best surgical candidates are those patients with focal, unilateral brain lesions (eg, mesial temporal sclerosis, tumor, or vascular malformation) that correlate with focal-onset seizures localizing to the ipsilateral temporal lobe. (See 'Mesial temporal lobe epilepsy' above.)

•Focal epilepsy with normal brain imaging – Patients with focal seizures of neocortical origin and normal brain MRI scans (nonlesional epilepsy) may also be candidates for epilepsy surgery, but management is often more challenging because of difficulty lateralizing and localizing the epileptogenic zone, difficulty defining the extent of brain tissue that must be resected for seizure freedom, and concerns regarding functional cortex. Seizure outcomes tend to be less favorable for these patients compared with patients who have mesial temporal lobe epilepsy. (See 'Focal epilepsy with normal brain MRI' above.)

•Lesional epilepsy – Low-grade primary brain tumors, vascular malformations, and malformations of cortical development (MCDs) are other causes of refractory focal epilepsy. Resection of the epileptogenic brain region often results in improved seizure control. Outcomes are typically better for patients with tumors and vascular malformations than for patients with focal cortical dysplasia. (See 'Lesional epilepsy' above.)

●Complications of surgery – Epilepsy surgery is associated with low rates of morbidity and mortality when performed at specialized centers. The most common adverse effects related to epilepsy surgery are cognitive impairment and visual field defects. (See 'Surgical complications' above.)