INTRODUCTION — Patients with epilepsy whose seizures do not successfully respond to antiepileptic drug (AED) therapy are considered to have drug-resistant epilepsy (DRE). This condition is also referred to as intractable, medically refractory, or pharmacoresistant epilepsy. As many as 20 to 40 percent of patients with epilepsy (roughly 400,000 Americans) are likely to have refractory epilepsy. The annual cost for patients with epilepsy in the United States is estimated to be approximately 12.5 billion dollars (based on a 1995 survey); DRE contributes a substantive proportion of this cost [1,2]. People with DRE have the greatest burden of epilepsy-related disabilities, further contributing to the scope of this problem.
Because of the need to individualize therapy, no rigid set of guidelines can be applied to determine medical intractability, however, population-based studies have provided information regarding the prognosis of DRE that are helpful in making treatment decisions. Resective surgical therapy for epilepsy has the potential to eliminate seizures in many patients with localization-related DRE.
This topic discusses the evaluation and approach to the management of individuals with DRE. Other issues regarding the evaluation and treatment of individuals with seizures and epilepsy are presented separately. (See "Overview of the management of epilepsy in adults" and "Evaluation of the first seizure in adults" and "Initial treatment of epilepsy in adults" and "Surgical therapy of epilepsy in adults" and "Vagus nerve stimulation therapy".)
DEFINITION — Traditionally, therapeutic failure of three antiepileptic drugs (AEDs) defined intractability [3-11]. With many new AEDs available in recent years, it might have been expected that more, rather than fewer, drug trials would be recommended before determining intractability. However, several prospective case series have shown that a high likelihood of medical intractability can be identified after two unsuccessful trials, as with each AED failure, the likelihood of successful treatment with other drugs diminishes [5-12]. As an example, in one study, 470 adolescent and adult patients with a diagnosis of epilepsy were started de novo on AED treatment and followed up to 16 years (median of five years). With the first drug trial, 47 percent became seizure-free. A second medication trial produced remission in 13 percent, and only 4 percent became seizure-free on a third medication regimen [5].
A task force of the International League against Epilepsy proposed that drug-resistant be defined as the failure of adequate trials of two tolerated, appropriately chosen and administered antiepileptic drugs (whether as monotherapy or in combination) to achieve seizure freedom [12]. They also recommended replacing the term “intractable” with “drug-resistant” epilepsy (DRE).
Some patients who meet this definition of DRE subsequently achieve prolonged (12 months or more) periods of seizure remission [13]. However, the risk of seizure relapse in these individuals remains high, greater than 70 percent in one series.
Frequency and severity of seizures are less commonly included in a definition of DRE [4,14]. These can vary among individuals with DRE and are important considerations when weighing treatment options. Similarly, it is important to understand the impact of seizures in the context of the individual's life, job, and other psychosocial circumstances [15]. Even infrequent seizures can have a large impact.
Less often considered, but also important, is the burden of adverse effects of AEDs that a patient experiences. If seizures can be controlled but only at medication doses that produce disabling side effects, then it may also be reasonable to consider that such a person has DRE.
EPIDEMIOLOGY — Because of unstandardized definitions as well as misdiagnoses, the incidence and prevalence of DRE are somewhat uncertain [16]. Estimates of the proportion of epilepsy cases that are or become medically resistant vary between 20 and 40 percent [3,11,17-20].
Risk factors — A number of prospective studies have attempted to identify factors that predict the risk of DRE. These studies have varied somewhat in their sampling (population-based versus hospital-based) and whether they included children, adults, or both [5]. No single factor has been found to be uniquely useful in making accurate predictions. A combination of one or several of these factors may help to define those who are not likely to respond to medical treatment.
- The response to the first AED trial is the most important and consistently cited predictive factor among both population-based and hospital-based studies and in both adult and pediatric populations [5,6,21-23]. While more than half of patients respond to the first AED prescribed, less than 20 percent are likely to respond to subsequent drug trials. AED failure due to lack of efficacy is a stronger predictor of DRE than failure due to intolerance or side effects [11]. With each number of failed AED trials, the risk of DRE increases [11,22,24-26].
- A high number of seizures prior to diagnosis and treatment is another consistently identified risk factor for DRE [5-7,11,17,21,27-32].
- The underlying etiology and seizure classification are also important. Idiopathic syndromes, for both generalized and localization-related epilepsy, have a better prognosis than symptomatic/cryptogenic epilepsy in both pediatric and adult populations [7,11,17,21,25,27,33-36].
Certain pediatric epilepsy syndromes are almost invariably medically intractable. These include early (neonatal) myoclonic encephalopathy, early infantile epileptic encephalopathy, Lennox Gastaut syndrome, and Rasmussen encephalitis among others.
Nonidiopathic localization-related epilepsy underlies more than half of the cases of DRE in adults [7]. Among the localization-related epilepsies, those associated with vascular lesions may be more treatment responsive than those with mesial temporal sclerosis (MTS), cortical dysgenesis, or dual pathology [6,33-35,37]. Individuals with MTS have some of the highest rates of medical intractability, 40 to 80 percent [3,6,8,34]. - Other findings more variably associated with the risk of DRE include a presentation with status epilepticus [7,21,24,33,36], a longer duration of epilepsy [25,27], a family history of epilepsy [11,17,29], a history of febrile convulsions [29], and abnormal EEG findings [7,11]. An abnormal neurologic examination and/or developmental delay are also identified as risk factors for DRE in some studies [17,24,28,33].
- Some studies, but not all, suggest that age at presentation may be a factor in the development of DRE [6,11,24,32,33]. Some pediatric studies have found that seizure onset in later childhood or adolescence appears to be more likely to be associated with DRE than seizures with onset between the ages of 5 and 10 years [7,28,36]. Onset in the neonatal time period has been associated with DRE in at least one series [28]. Individuals who develop epilepsy later in life (>65 years) appear to be less likely to develop DRE than younger adults [20,38]. The varying risk of DRE by age group likely relates to the underlying pathogenesis of epilepsy that also varies by age.
The studies cited above have generally been performed in studies examining a patient's response to initial AED trials. However, a smaller percentage of patients with epilepsy enter remission early in their course and develop DRE later, after a period of remission that in some cases is a long as several years [9,20]. Factors that predict a later development of DRE are not well defined. However, this phenomenon of delayed intractability is most commonly described in the setting of mesial temporal sclerosis. (See 'Pathogenesis' below.)
PATHOGENESIS — It is uncertain why seizures are or become medically resistant in any given individual.
Medical intractability may be a feature of epilepsy at the time of presentation or may evolve over time. Prospective studies with a long duration of follow-up suggest that 70 to 80 percent of patients retain their status as intractable versus in remission [4,9,17,20,21,31,39-41]. This means that a minority of patients with apparent DRE are subsequently able to achieve remission and also that some patients whose seizures are initially controlled will later relapse. In the latter group, medical control can be regained with subsequent drug trials in some patients, while in others, medical intractability is then permanent [42-44]. These different clinical courses may represent different mechanisms of medical intractability or pharmacoresistance.
A delayed development of intractability is most often described with epilepsy onset in childhood, particularly with epilepsy associated with mesial temporal sclerosis (MTS) [9,45,46]. Accumulating evidence suggests that in some individuals, MTS is a progressive condition. Pathologic studies demonstrate an evolving process involving glial proliferation and dendritic sprouting with synaptic reorganization [11,47]. Longitudinal neuroimaging studies also support a progressive process [32,48]. Alterations in neural circuitry conceivably may lead to an epileptic network that becomes drug resistant over time [49]. (See "Localization-related (partial) epilepsy: Causes and clinical features" and "Pathophysiology of seizures and epilepsy", section on 'Consequences of repeated seizures'.)
Studies in animal models comparing drug-resistant and drug-sensitive temporal lobe epilepsy have found that the former is associated with altered expression of multidrug transporters, altered expression of AED targets, as well as morphologic alterations in the hippocampus [26]. Distinguishing which of these findings is the cause versus the effect of IE, and if it has relevance in human IE, are the subjects of ongoing investigations. As an example, one study found an association with a polymorphism in the drug transporter gene (ABCB1 or MDR1) and drug-resistant epilepsy [50], although this was not corroborated in follow-up studies [51-53]. Another hypothesis proposes that overexpression of these multidrug transporters is induced by recurrent seizures, potentially providing an impetus for early aggressive seizure treatment [26,54]. Other alterations, acquired or inherited, of AED absorption, metabolism, receptor-binding, and blood brain barrier permeability are also potential causes of pharmacoresistance [55,56].
COMPLICATIONS OF INTRACTABLE EPILEPSY — Individuals with DRE have an increased mortality rate, estimated at 1.37 per 100 person-years [23,57-59]. The standardized mortality ratio for patients with recurrent seizures is 4.69 [58]. Individuals who become seizure free have no increased mortality [60]. Some deaths are related to the underlying cause of epilepsy (eg, cerebral neoplasm, neurodegenerative disease); other deaths are directly seizure-related, such as those that occur in the context of status epilepticus and in seizure-related accidents. Sudden unexplained death in epilepsy patients (SUDEP) is 40 times more likely among patients who continue to have seizures than in those who are seizure free [57]. (See "Sudden unexpected death in epilepsy" and "Overview of the management of epilepsy in adults", section on 'Mortality'.)
Nonfatal injuries are also common in those with DRE. These include head injury, burns, and fractures, among others; most are seizure-related [61].
IE is also associated with disability and diminished quality of life [8,11,62]. Examples include poor academic achievement, unemployment, and social isolation. Most patients with DRE cannot drive. In one community-based survey of people with epilepsy, those with self-reported incomplete seizure control were more likely to express concerns about the feelings of fear, their quality of life, work, adverse effects of therapy, and the stigma associated with their condition, even when their seizures were relatively infrequent [63]. These complications of DRE result from the combined effects of recurrent seizures, AED toxicity, comorbid depression, as well as psychosocial factors such as excessive dependency [8,11,62]. (See "Overview of the management of epilepsy in adults", section on 'Psychosocial issues'.)
DIFFERENTIAL DIAGNOSIS — When a patient's seizures do not appear to respond to AED therapy, the clinician should reconsider the seizure classification and the appropriateness of the AED regimens that have been employed. Clinicians should also reconsider the diagnosis of epilepsy. Misdiagnosis is common; in one series, as many as 26 percent of individuals thought to have DRE were incorrectly diagnosed most often as a result of incomplete history-taking and/or EEG misinterpretation [64].
Apparent intractability — It is important to differentiate true versus apparent medical intractability. Reasons for apparent treatment failure that do not reflect intractability include:
- An incorrect diagnosis of seizure classification leading to incorrect drug choice. It is not uncommon for idiopathic generalized epilepsy syndromes to be unrecognized and inappropriately treated with AEDs that are more appropriate to localization-related epilepsy (table 1) [3]. In some instances, a narrow-spectrum AED can worsen seizure frequency in individuals with generalized epilepsy. One example is when carbamazepine is prescribed for juvenile myoclonic epilepsy [64-66].
- Inappropriate dose. In other cases, inadequate dosing or frequency of AED dosing has led to apparent intractability [64]. In contrast, seizures can also occur with AED toxicity. This has been described in rare cases in association with phenytoin, carbamazepine, tiagabine, and valproate [66].
- Noncompliance. Compliance with AED is frequently imperfect. In one case series, 71 percent of patients reported at least occasional dose omissions, and 45 percent reported a seizure after a missed dose [67]. Providing a nonjudgmental setting in which to elicit this history is important. Support from family members and physicians increases medical compliance [3,68,69]. (See "Overview of the management of epilepsy in adults", section on 'Noncompliance with AED therapy'.)
- Lifestyle factors. Examples of lifestyle factors that can increase seizure frequency include recreational drug or alcohol abuse and sleep deprivation [3,29,69].
Psychogenic nonepileptic seizures — Psychogenic nonepileptic seizure (PNES) can mimic epileptic seizures. In contrast to epileptic seizures, PNES are not associated with physiological central nervous system dysfunction but are instead psychogenically determined. PNES typically do not respond to AED therapy. While not without limitations, video-EEG monitoring is the gold standard test for the diagnosis of PNES.
Among patients referred to epilepsy monitoring units for apparent IE, 25 to 40 percent are diagnosed with PNES [70,71]. The clinical features, diagnosis, and treatment of PNES are discussed separately. (See "Psychogenic nonepileptic seizures".)
Other nonepileptic paroxysmal disorders — In addition to PNES, other nonepileptic paroxysmal events, especially syncope but also certain sleep and movement disorders, can be mistaken for epilepsy (table 2) [64,72]. (See "Nonepileptic paroxysmal disorders in adolescents and adults".)
EVALUATION — Patients with DRE should have further testing to confirm the diagnosis of epilepsy and also to better define the epilepsy syndrome and underlying classification in order to best direct treatment [15,73]. In most cases, the evaluation of DRE will include video-EEG monitoring and magnetic resonance imaging [69].
Video EEG monitoring — Inpatient video-EEG monitoring combines both a video and EEG recording of clinical events. This test is used primarily to determine whether epilepsy is the cause of recurrent seizure-like events. In some series, more than 25 percent of individuals referred for monitoring for refractory epilepsy are found to have nonepileptic events, usually psychogenic nonepileptic seizures [64]. EEG monitoring can also aid in seizure classification and is used for presurgical evaluation of epilepsy patients. (See "Video and ambulatory EEG monitoring in the diagnosis of seizures and epilepsy".)
Neuroimaging — By the time a patient is considered to have IE, a magnetic resonance imaging (MRI) study will usually have been performed. In many cases, this should be repeated, particularly if the original study was unrevealing. The sensitivity of MRI for an underlying cause of epilepsy (so-called lesional epilepsy) can be substantially improved by using an epilepsy protocol; these are not routinely used outside of subspecialty epilepsy centers. (See "Neuroimaging in the evaluation of seizures and epilepsy", section on 'Sensitivity'.)
Not all MRI findings are relevant; isolated findings of diffuse atrophy, punctate foci of T2 signal abnormalities in the white matter, and other nonspecific findings are not known to be epileptogenic. MRI findings should be correlated with the patient's seizure semiology and EEG findings; some potentially epileptogenic lesions may be incidental.
In the absence of a causative lesion on MRI, an epileptogenic focus can sometimes be defined in patients with localization-related epilepsy using advanced neuroimaging techniques including positron emission tomography (PET), single photon emission computed tomography (SPECT), and magnetic source imaging (MSI). The choice of study often depends upon the availability and expertise at a particular center. The use of these tests is discussed in detail separately. (See "Surgical therapy of epilepsy in adults" and "Neuroimaging in the evaluation of seizures and epilepsy".)
Seizure diaries — In some patients, it may be helpful to have them carefully record seizures, along with other relevant information, including dietary changes, timing of medication intake of both AEDs and other drugs, amount and quality of sleep, and menstrual cycle changes. This may serve to improve compliance and may also help in identifying precipitants, such as hormonal changes associated with the menstrual cycle. (See "Catamenial epilepsy".)
TREATMENT OPTIONS — Resective epilepsy surgery is the treatment of choice for medically resistant lesional partial epilepsy as this has the most likely chance of producing remission. Further AED trials, vagal nerve stimulation, and the ketogenic diet can reduce seizure frequency and improve quality of life but are more likely to be palliative, rather than curative, treatment options [15].
Antiepileptic drugs — Further medications trials of AEDs in mono- or polytherapy can be of benefit in individuals with epilepsy. It is important to review past treatment trials with the patient to assess whether the dose or frequency of dosing was adequate. The appropriateness of past AED trials to the individual’s seizure-type (table 1) should also be specifically evaluated, as should compliance and any potential barriers to compliance. (See "Overview of the management of epilepsy in adults", section on 'Antiepileptic drug therapy'.)
Sequential drug trials have a small likelihood of inducing remission in patients who have already failed two or more AED regimens. This approach can produce remission rates estimated at 4 to 6 percent per year, or a cumulative rate of 14 to 20 percent [24,25,27,33,35,42]. Among those who do not become seizure-free, a substantial reduction in seizure frequency is possible; in different series, 21 to 70 percent of patients achieve a 50 percent or greater reduction in seizure frequency [3,25,27,33]. Reduction in seizure severity may also improve patients’ quality of life [74]. However, studies with long-term follow-up find that the benefit of successive drug trials is not sustained in one-fourth or more [42,43].
Choosing an AED with a different mechanism of action than one not previously efficacious may maximize the benefit from subsequent drug trials (table 3). Some suggest that drug combinations employing AEDs with different mechanisms of action may also be fruitful, although neither of these approaches has been systematically evaluated. An analysis of 70 randomized controlled trials of AEDs administered as add-on therapy in patients with refractory partial epilepsy found that differences in efficacy were of two small a magnitude to allow conclusions about which AED is more effective in this setting [75]. The approach to successive drug trials is discussed in more detail separately. (See "Overview of the management of epilepsy in adults", section on 'Subsequent drug trials'.)
Enrollment in clinical drug trials for investigational AEDs is a treatment option for patients who do not wish to proceed with current therapy, are not surgical candidates, or whose preference is to continue with medical management. Epilepsy centers and other neurologic centers often have several active ongoing research protocols.
Epilepsy surgery — Epilepsy surgery should be considered in appropriate patients with DRE when seizures are sufficiently frequent or severe as to significantly disrupt the patient's quality of life. These considerations are necessarily individualized, but in general include those seizures that impair consciousness, cause injury, and occur sufficiently frequently as to be disabling [76].
Resective epilepsy surgery has the best-established efficacy for individuals with lesional temporal lobe epilepsy [77,78]. Patients with concordant abnormalities in one temporal lobe on MRI and EEG have a rate of seizure remission as high as 90 percent [79,80]. Patients with nonlesional temporal lobe epilepsy also have a high remission rate with surgical therapy. The efficacy is highest in patients in whom EEG and another imaging modality (eg, SPECT, PET, or MSI) reveal a consistent location of the epileptic focus.
Neocortical focal epilepsy also responds to resective surgery. As with mesial temporal lobe epilepsy, rates of seizure remission are highest in patients who have MRI lesions that are concordant with the anatomic focus of seizure activity on EEG. However, localization using SPECT, PET, and/or MSI, can also define a seizure focus that when surgically removed leads to seizure remission rates that exceed 50 percent.
Other surgical treatments (lobar and multi-lobar resections, hemispherectomy, corpus callosotomy, multiple subpial transections) are sometimes employed for palliative treatment in children and sometimes adults with catastrophic epilepsy syndromes.
Epilepsy surgery is discussed in more detail separately. (See "Surgical therapy of epilepsy in adults" and "Treatment of seizures and epileptic syndromes in children", section on 'Epilepsy surgery'.)
Vagus nerve stimulation — Vagus nerve stimulation (VNS) has been approved for adjunctive treatment of medically intractable partial onset seizures in adults and children. Approximately 30 to 40 percent of patients achieve a greater than 50 percent reduction in seizure frequency, a benefit that is sustained over time [81-84]. Serious adverse events are rare.
VNS is a valid treatment option for patients with well-documented IE, who are either opposed to intracranial surgery, or who are not candidates for intracranial surgery, or whose seizures were not substantially improved by prior intracranial epilepsy surgery [85-87]. Resective surgery for appropriate candidates is preferred over VNS because of the substantially greater potential for complete seizure remission.
VNS is discussed in detail separately. (See "Vagus nerve stimulation therapy".)
Other stimulation approaches — The use of brain implants that provide focal electrical stimulation within the brain as a means of controlling DRE is under investigation.
- Deep brain stimulation — Subcortical stimulation paradigms have targeted the anterior and centromedian thalamic nuclei, the subthalamic nucleus, the caudate, and the cerebellum [88,89]. Open-label and some small controlled studies have found that stimulation in these sites appears to reduce seizure frequency by 50 percent or more in some patients [88-91].
In a randomized clinical trial of stimulation in the anterior nucleus of the thalamus (SANTE trial) in 110 patients with drug-resistant epilepsy, stimulation therapy was associated with a 29 percent reduction in seizure frequency compared with controls; 54 percent of patients had a seizure reduction of at least 50 percent. Complex partial and “most severe” seizures were most significantly reduced by stimulation. Participants in the stimulated group were more likely to report depression and memory problems as adverse affects [92]. Despite the positive results of the trial, the United States Federal Drug Administration deferred approval of the device and technique until further trials were completed. The European Union has approved use of the device and technique for refractory epilepsy in its member countries. - Cortical stimulation — Two small controlled trials, as well as other open label studies, have found that stimulation of the hippocampus appears to reduce seizure frequency in patients with mesial temporal lobe epilepsy [93-96].
Another approach under investigation employs a closed-loop cortical stimulation unit coupled to a seizure-detection system [88,89,97,98]. Open-label studies have found that this treatment may be associated with a substantial reduction in seizure number, intensity, and duration. In a clinical trial, 191 adults with medically intractable epilepsy were randomly assigned to sham or active stimulation in response to seizure detection [99]. After 12 weeks of therapy, the reduction in seizure frequency greater in the active compared to the sham treatment arm (37.9 versus 17.3 percent). Seizure reduction was sustained over a subsequent 84 week open label period in which all patients received active stimulation. Active treatment was also associated with improved quality of life; neither mood nor cognition was negatively impacted. - Transcranial magnetic stimulation — Low frequency transcranial magnetic stimulation also reduces cortical excitability [100,101]. Uncontrolled trials and case reports have suggested that this may reduce seizure frequency [102,103]. However, small controlled trials have had mixed results [104,105].
Ketogenic diet — The ketogenic diet (high-fat, low protein) diet has demonstrated efficacy in children with IE, with more than one-third experiencing a 50 percent or greater reduction in seizures
In two small case series of adult patients, the traditional ketogenic diet and a modified Atkins diet reduced seizure frequency by 50 percent or more in half of patients with DRE [106,107].
The ketogenic diet is discussed in detail separately. (See "The ketogenic diet".)
Catamenial seizures — Women with catamenial epilepsy may benefit from specific interventions to lower seizure frequency. (See "Catamenial epilepsy".)
SUMMARY AND RECOMMENDATIONS — It is estimated that between 20 to 40 percent of patients with epilepsy will not have complete seizure control with antiepileptic drug (AED) therapy alone.
- Prospective studies indicate that most patients with drug-resistant epilepsy (DRE) can be identified early in their presentation, after a failure of two AED trials. (See 'Definition' above.)
- Predictors of DRE include lack of efficacy of a first AED trial, a high number of seizures prior to treatment, and a symptomatic/cryptogenic rather than idiopathic epilepsy syndrome. (See 'Risk factors' above.)
- Individuals with DRE have an increased risk of mortality as well as other disabilities including poor academic performance, unemployment, and other lifestyle restrictions. This provides an impetus for aggressive treatment. (See 'Complications of intractable epilepsy' above.)
- Patients with DRE should undergo evaluation (usually video EEG monitoring) to confirm the diagnosis of epilepsy; as many as 20 percent of patients with apparent DRE will have a nonepileptic paroxysmal disorder, usually psychogenic nonepileptic seizures. EEG monitoring can also aid in seizure classification; findings should be correlated with the clinical history. (See 'Evaluation' above.)
- Patients with localization-related DRE should have a magnetic resonance imaging study (MRI) to identify a potential surgical lesion. The sensitivity of MRI can be enhanced by the use of an epilepsy MRI protocol. Other neuroimaging studies (eg, SPECT, PET) can be employed in individuals in whom MRI does not show a lesion, in whom the MRI lesion does not correspond to the EEG localization, or who have dual pathology. (See 'Neuroimaging' above.)
- We recommend surgical evaluation for patients with localization-related or partial epilepsy (Grade 1A). Magnetic resonance imaging (MRI) scan performed using an epilepsy protocol will often identify a lesion (eg, mesial temporal sclerosis, cortical dysplasia) amenable to surgical resection. The best efficacy for surgery is achieved in individuals with concordant localization of MRI and EEG abnormalities. For patients without a MRI lesion, an MRI lesion that is discordant with EEG findings, or with dual pathology, further neuroimaging studies can help identify an epileptic focus that might respond to surgical therapy. (See "Surgical therapy of epilepsy in adults" and "Neuroimaging in the evaluation of seizures and epilepsy".)
- For patients in whom epilepsy surgery is not an option or whose seizures persist after surgery, we suggest treatment trials with other AEDs appropriate for their epilepsy syndrome and/or vagus nerve stimulation (Grade 1A). While the chance of seizure remission with these treatments is not high, reductions in seizure frequency and improved quality of life are possible in most. (See "Vagus nerve stimulation therapy" and 'Antiepileptic drugs' above.)
Searching for
