INTRODUCTION — Most children with epilepsy achieve reasonably good seizure control with antiseizure medication therapy, but some are refractory despite numerous medications. Medical treatment failure is often apparent early in the course of treatment. In these cases, referral to a comprehensive epilepsy center is appropriate to explore additional therapeutic options, including epilepsy surgery, vagus nerve stimulation, and the ketogenic diet.
There is no standardized definition of medically intractable epilepsy. A task force of the International League Against Epilepsy proposed that drug-resistant epilepsy be defined as failure of adequate trials of two tolerated and appropriately chosen and used antiseizure medication schedules (whether as monotherapies or in combination) to achieve sustained seizure freedom [1].
This topic discusses the management of seizures and epilepsy in children who are not controlled with initial antiseizure medication therapy. The clinical features, diagnosis, and initial management of seizures and epilepsy in children are presented separately:
Seizures and epilepsy in children: Classification, etiology, and clinical features
Seizures and epilepsy in children: Clinical and laboratory diagnosis
Seizures and epilepsy in children: Initial treatment and monitoring
Epilepsy syndromes in children
Epilepsy in children: Comorbidities, complications, and outcomes
EVALUATION — When seizures do not respond as expected to initial antiseizure medication therapy, it is important to reconsider the seizure classification and the appropriateness of the antiseizure medication regimens that have been tried. In addition, clinicians should consider whether the diagnosis of epilepsy is accurate, as misdiagnosis is common. Common mimics of seizures include psychogenic nonepileptic seizures and other nonepileptic paroxysmal disorders (table 1).
Additional reasons for apparent treatment failure that do not reflect true intractability include:
●An incorrect diagnosis or seizure classification (ie, focal versus generalized), leading to an inappropriate antiseizure medication choice that might be aggravating seizures (table 2). (See "Seizures and epilepsy in children: Initial treatment and monitoring", section on 'Seizure-related considerations'.)
●Inappropriate antiseizure medication dosing (either too high or too low) or frequency. (See "Seizures and epilepsy in children: Initial treatment and monitoring", section on 'Drug administration and dosing'.)
●Nonadherence to antiseizure medication therapy. (See "Seizures and epilepsy in children: Initial treatment and monitoring", section on 'Adherence to antiseizure medications'.)
MEDICAL THERAPY — Children who fail to respond to antiseizure medication monotherapy at adequate doses or do not tolerate effective doses should be started on a second antiseizure medication, although the likelihood of complete seizure remission decreases with each subsequent failed antiseizure medication trial. The likelihood of success varies based on the underlying etiology.
Adding a second antiseizure medication — The first antiseizure medication fails in 20 to 40 percent of children with epilepsy; lack of efficacy and side effects contribute roughly equally to treatment failure [2]. Adding a second antiseizure medication is a reasonable next step when seizures are resistant to adequate doses of the initial drug.
●If the initial drug was partially effective, it should be continued until reasonable levels of the new antiseizure medication are achieved. Tapering the first drug can then be attempted if seizures are controlled.
●If the initial antiseizure medication is ineffective, it can be tapered earlier, as the dose of the second drug is increased. The pharmacokinetics of either drug can be altered by the other; serum levels may have to be monitored closely.
A second antiseizure medication may also be considered in children with several different seizure types when monotherapy is not effective. In one practice, the clinicians found that children with status epilepticus, developmental disabilities, and multiple seizure types were more likely to require polytherapy than those without these features [3].
Virtually every combination of antiseizure medications has been used. Certain combinations should be avoided when mechanisms of action overlap and toxicities are additive, such as any combinations of phenobarbital, primidone, and benzodiazepines, all of which are central nervous system depressants. Phenytoin and carbamazepine also have overlapping mechanisms of action and are usually not given together. (See "Antiseizure medications: Mechanism of action, pharmacology, and adverse effects".)
There are instances in which the synergistic interaction of two antiseizure medications improves seizure control. Children with certain generalized epilepsy syndromes occasionally require a combination of valproate and ethosuximide or lamotrigine. (See "Childhood absence epilepsy", section on 'Treatment' and "Juvenile myoclonic epilepsy", section on 'Adjunctive therapy'.)
Inadequate response to multidrug therapy — There are numerous reasons why seizure control may be inadequate despite multidrug therapy, and in some cases seizure frequency is improved when the number of antiseizure medications is reduced:
●In many cases, a second antiseizure medication was added to an initial, ineffective antiseizure medication that was not then discontinued. If the seizures increased, a third drug often was added. The first or second antiseizure medication may actually have increased the seizure frequency, and the problem was compounded further with the addition of more antiseizure medications.
●Most of the antiseizure medications have been reported to increase seizure frequency in a small percentage of patients and, in some instances, induce new seizure types [4]. These seizures may occur at low or therapeutic doses of some drugs and at toxic doses in others. Carbamazepine, for example, can precipitate atonic seizures, particularly in children with Lennox-Gastaut syndrome. Phenytoin can worsen absence seizures at therapeutic doses and cause generalized seizures at toxic doses.
Thus, the initial approach to the child with intractable seizures on polytherapy is to slowly decrease the antiseizure medication that has been the least effective.
When a child has intractable epilepsy despite treatment with more than one antiseizure medication, it is also reasonable to reevaluate the diagnosis of epilepsy and initiate referral to a comprehensive epilepsy center to determine eligibility for epilepsy surgery and other treatment modalities, such as dietary therapy and vagus nerve stimulation [5]. (See 'Epilepsy surgery' below and 'Vagus nerve stimulation' below and 'The ketogenic diet' below.)
Home rescue therapy (transmucosal antiseizure medications) — Parents and caregivers of children with frequent, prolonged seizures are often provided with transmucosal benzodiazepines for home rescue therapy in an attempt to overcome delays in treatment related to emergency transportation and provision of intravenous access. Such problems have led to increased use of transmucosal routes of antiseizure medication delivery, such as rectal, buccal, intranasal, and sublingual administration, that allow for earlier initiation of rescue therapy in the home or prehospital setting [6].
The largest experience is with rectal diazepam, although accumulating evidence suggests that buccal midazolam is also safe and may be more effective than rectal diazepam [6-8]. Intramuscular and intranasal administration of midazolam appears to have similar efficacy as rectal diazepam. Transmucosal administration of midazolam will likely become first-line therapy for out-of-hospital emergency therapy once a convenient formulation for nonmedical personnel is approved and available. (See 'Rectal therapy' below and 'Buccal therapy' below.)
Prolonged seizures that do not respond to transmucosal therapy require hospitalization, intravenous administration of antiseizure medications, and may require intensive care unit management. This is discussed separately. (See "Management of convulsive status epilepticus in children".)
Rectal therapy — Rectal benzodiazepines are an effective and relatively safe emergency treatment, particularly for patients who live at some distance from a medical facility, those requiring emergency therapy while traveling, and those whose seizures occur in clusters.
Diazepam is available in a rectal gel (Diastat) that is obtained from the pharmacy in prefilled ready-to-use syringes that are set to deliver a specific dose. These have a long shelf life (years) and do not require refrigeration. The rates of absorption and response are slower than those of intravenous diazepam but faster than with oral administration. Respiratory depression is extremely rare [9], although if concerns about this possible complication exist, particularly in a handicapped child, a test dose can be given under physician supervision.
Rectal paraldehyde may also be used in the setting of a prolonged convulsion when other routes of antiseizure medication administration are not available. It is administered as a dose of 0.8 mL/kg of a 0.4 mL/kg liquid mixed 1:1 with olive oil. In a case series of 53 episodes of prolonged convulsions in 30 patients, rectal paraldehyde terminated the convulsion in 62 percent [10]. There was no respiratory depression or other adverse events recorded. Paraldehyde is not available in the United States.
Many antiseizure medications can be given as a rectal suppository for short periods (eg, postoperatively) and occasionally as chronic treatment for patients unable to take oral medications. Any antiseizure medication available as an oral suspension (clonazepam, carbamazepine, valproate, oxcarbazepine, levetiracetam) can be given rectally for short periods. Carbamazepine has a significant cathartic effect.
Buccal therapy — Buccal midazolam therapy uses the intravenous formulation of midazolam hydrochloride delivered into the buccal cavity between the gum and cheek [7]. Buccal midazolam is approved for treatment of acute seizures in children in the United Kingdom and the European Union but is not yet approved in the United States. The suggested doses are 2.5 mg for infants between 6 months and 1 year, 5 mg for those aged 1 to <5 years, 7.5 mg for those aged 5 to <10 years, and 10 mg for those age 10 to <18 years [6].
Buccal midazolam is effective for the emergency treatment of seizures in children, as demonstrated by at least three randomized clinical trials [11,12]. In the earlier and smaller trial that evaluated 79 seizures in 28 children with severe epilepsy, buccal midazolam was as effective as rectal diazepam at stopping seizures [11].
A subsequent multicenter randomized controlled trial evaluated 219 seizures in 177 children ages six months and older (median age three years) with new-onset seizures or existing epilepsy who presented to the emergency department with ongoing seizures and no established intravenous access [12]. Buccal midazolam was more effective than rectal diazepam for either stopping seizure activity within 10 minutes without respiratory depression or preventing another seizure within one hour (56 versus 27 percent, respectively). The risk of respiratory depression with buccal midazolam was low and similar to the risk with rectal diazepam. The dose of buccal midazolam used was 2.5 mg, 5 mg, 7.5 mg, and 10 mg for children ages 6 to 12 months, 1 to 4 years, 5 to 9 years, and 10 or more years, respectively.
A randomized trial compared buccal midazolam with intravenous diazepam in 120 children presenting to an emergency department with convulsive seizures [13]. Time to treatment was shorter with buccal midazolam; time from administration to seizure control was shorter with diazepam. Overall, there was a shorter time from treatment decision to control of seizures for buccal therapy.
Intranasal therapy — A nasal spray formulation of diazepam is available; it is approved by the US Food and Drug Administration (FDA) for the acute treatment of intermittent, stereotypic episodes of frequent seizure activity (ie, seizure clusters, acute repetitive seizures) in patients age 6 years and older [14].
Nasal midazolam is also approved by the FDA for the acute treatment of intermittent, stereotypic episodes of frequent seizure activity (ie, seizure clusters, acute repetitive seizures) in patients age 12 years and older [15]. A 2021 systematic review and meta-analysis of 10 studies comparing intranasal midazolam (n = 169) versus intravenous or rectal benzodiazepines (n = 161) for the treatment of acute seizures found no significant difference between the two treatment groups for time to seizure cessation or control of seizures [16].
Intranasal lorazepam also appears to be an effective treatment option [17], and in one study appeared to have similar efficacy to intravenous lorazepam [18].
School action plans — A school action plan should be in place for children with frequent, prolonged seizures who use rescue therapy in the home environment, with specific instructions for medical and nonmedical school personnel [19,20]. The American Academy of Pediatrics has published a guideline highlighting practical and legal considerations for clinicians, families, guardians, caregivers, and schools when developing school-based plans [21]. In addition to parameters about when a rescue medication should be given, tailored to the child's individual history, plans should also generally include discussion of potential adverse effects, guidance about the types of situations in which personnel should seek further medical assistance, and discussion of when a child is safe to remain in school after a seizure.
EPILEPSY SURGERY — Surgical interventions should be considered, regardless of age, in children who have persistent, frequent seizures that are having an adverse impact upon their lives or are interfering with their cognitive and psychosocial development [22,23]. The type of surgery that a child may be eligible for is individualized based on a comprehensive epilepsy evaluation.
Surgical evaluation — Eligibility for epilepsy surgery is determined by a comprehensive evaluation that includes efforts to identify the epileptogenic zone and a determination of the extent to which it can be resected safely. Components of the evaluation include a clinical evaluation, neuropsychologic testing, routine and video electroencephalography (EEG), high-resolution magnetic resonance imaging (MRI), and other advanced imaging techniques. Components of the surgical evaluation of epilepsy are reviewed in detail separately. (See "Surgical treatment of epilepsy in adults", section on 'Surgical evaluation'.)
Specific procedures — Surgical approaches can be divided into relatively less invasive procedures, such as vagus nerve stimulation, and surgeries such as focal resections, lobar or multilobar resections, corpus callosotomy [24], hemispherectomy, and multiple subpial transection [25].
The procedure selected depends upon the type and localization of the seizures. Children with a well-localized epileptic focus, particularly if a lesion is present on imaging studies in the same location, do very well with removal of the abnormal tissue (lesionectomy). Children with uni-hemispheric syndromes such as hemimegalencephaly [26,27], Sturge-Weber syndrome, or Rasmussen disease [28,29] can be candidates for removal (hemispherectomy) or undercutting of the cortex of an entire hemisphere (functional hemispherectomy). An individualized approach is taken for complex lesions such as polymicrogyria, in which the epileptic zone may only partially overlap with the structural abnormality [30,31].
When patients are appropriately selected for epilepsy surgery after comprehensive surgical evaluation, observational studies and a single randomized trial in children find that seizure outcomes are superior to those achieved with continued medical therapy alone [32,33]. Improvements in seizure frequency are generally highest after medial temporal lobectomy or extratemporal lesional surgeries, intermediate after hemispherectomy, and lowest after corpus callosotomy. The risk of new or worsened neurologic deficits such as hemiparesis varies by specific procedure and preoperative neurologic baseline and is generally highest with hemispherectomy and extratemporal resections and lowest with medial temporal lobectomy.
In the only randomized trial of epilepsy surgery in children to date, one-year seizure freedom rates after surgery and appropriate medical therapy were superior to wait-list control and medical therapy alone for the entire cohort (77 versus 7 percent; n = 116) and varied according to the performed or planned procedure [32]:
●Temporal lobectomy (n = 29) – 100 versus 13 percent
●Extratemporal resection (n = 31) – 92 versus 5 percent
●Hemispherectomy (n = 23) – 87 versus 0 percent
●Corpus callosotomy (n = 26) – 0 versus 6 percent
●Hypothalamic hamartoma resection/disconnection (n = 7) – 100 versus 0 percent
Clinicians, patients, and parents were necessarily unblinded. Nearly all secondary outcomes improved in the surgical group at one year compared with the presurgical baseline, including scores on scales of seizure severity, quality of life, social maturity, and child behavior. Hemiparesis was the most common severe adverse event and occurred exclusively in the hemispherectomy group (26 percent); all but two patients recovered to antigravity strength or better by 12 months.
Temporal lobectomy — The most commonly performed invasive surgical procedure for epilepsy is temporal lobectomy as a treatment for focal-onset seizures arising from the medial temporal lobe. Among adults, the most common histopathologic diagnosis in temporal lobectomy specimens is hippocampal sclerosis [34,35]. Among children undergoing epilepsy surgery, hippocampal sclerosis makes up only approximately 15 percent of cases; more commonly, children have cortical malformations or low-grade tumors in temporal or extratemporal locations [35,36]. (See 'Extratemporal cortical resection' below.)
Cohort studies and a single randomized trial suggest that children who undergo temporal lobectomy for mesial temporal lobe epilepsy can achieve similar, if not greater, benefits on seizure control compared with adults, with reported rates of complete seizure freedom ranging from 65 to 100 percent at one to five years after surgery [32,36-39].
Extratemporal cortical resection — The results of resective surgery in children with extratemporal foci are similar to those of temporal lobectomy if a discrete lesion is present on preoperative neuroimaging. This appears to be true even if preoperative EEG recording demonstrates bilateral or generalized epileptiform activity [40]. Malformations of cortical development and low-grade primary brain tumors (eg, ganglioglioma, dysembryoplastic neuroepithelial tumor) are the most common causative lesions (both temporal and extratemporal) [35].
Rates of complete seizure remission range from 80 to 92 percent at one year after extratemporal cortical resection of lesional tissue [32,35,41,42].
Results in nonlesional cases (MRI-negative) are less gratifying, with success rates reported as less than 40 percent in one series [43]. However, the application of newer technologies in preoperative assessment may modify this somewhat. One study found that surgical success rates similar to those seen with MRI-positive patients were seen in MRI-negative patients in whom the epileptogenic region was identified by combination of ictal single-photon emission computed tomography (SPECT), positron emission tomography (PET), and intracranial EEG monitoring [44].
The occurrence of early (within two weeks) postoperative seizures was seen in 26 percent of children in one study and identified patients less likely to have a long-term seizure remission [45].
Cohort studies suggest that individuals with unifocal lesions, particularly low-grade tumors, are more likely to achieve seizure remission than those with no or multifocal lesions and/or cerebral dysplasia [35,38,39].
Hemispherectomy — Hemispherectomy is performed in children whose seizures are associated with a disease that diffusely affects one cerebral hemisphere. (See "Focal epilepsy: Causes and clinical features", section on 'Hemispheric syndromes'.)
The most common causes of seizures in children undergoing hemispherectomy include encephalomalacia from a remote insult such as perinatal stroke, hemimegalencephaly, multilobar cortical dysplasia, Rasmussen encephalitis, and Sturge-Weber syndrome.
Seizure control rates after hemispherectomy are generally favorable across a range of etiologies, with approximately two-thirds of patients achieving seizure freedom at five years [46-53]. Predictors of worse seizure outcomes in most but not all studies include hemimegalencephaly [46,51,54,55], bilateral fluorodeoxyglucose (FDG)-PET abnormalities [50], bilateral MRI abnormalities [54], longer duration of epilepsy preceding surgery [47,56], and early postoperative seizure [45,50]. Of these, duration of epilepsy prior to surgery is the most consistently reported and underscores the importance of early surgical evaluation in patients with refractory seizures.
Contralateral MRI abnormalities are a common cause for concern in selecting patients for hemispherectomy, but their importance is controversial. This is demonstrated by two studies with conflicting results:
●In a retrospective series of 110 patients who underwent hemispherectomy, 74 percent of cases had contralateral MRI abnormalities, the majority of which were described as mild to moderate, including nonspecific white matter changes [55]. At a median follow-up of two years, rates of seizure freedom were similar in those with and without contralateral MRI abnormalities (79 and 83 percent).
●A smaller study of 43 patients found contralateral MRI abnormalities (described as "unambiguous" findings on visual inspection) in 26 percent of patients who underwent hemispherectomy [54]. After a median follow-up of seven years, patients with contralateral MRI abnormalities had lower rates of seizure freedom (45 versus 88 percent).
The definition of a contralateral MRI abnormality has varied across studies, and this may partly explain the discrepancy in results. In addition, the length of follow-up often differs, and those studies with shorter follow-up times might overestimate seizure control rates in one or both groups.
Other imaging modalities may also be helpful in selecting patients who will optimally benefit from surgery. In the largest study to date, FDG-PET activity in the contralateral hemisphere was a stronger predictor of seizure control postoperatively than MRI abnormalities; seizure freedom was achieved in 72 percent of the 119 patients with unilateral PET abnormalities and only 44 percent of the 18 patients with bilateral abnormalities [50].
Motor and cognitive outcomes after hemispherectomy are not as well studied as seizure outcomes. Most patients already have moderate to severe hemiparesis prior to undergoing hemispherectomy. The risk of added motor deficits is approximately 25 percent in the immediate postoperative period; most children regain function to their preoperative baseline or better by one year [32].
Postsurgical developmental gains reported in most studies are modest, but development is often limited by severe preexisting brain damage that may involve the nonoperated hemisphere [49,54,56]. As with seizure control, the most consistent predictor of cognitive and motor outcomes is shorter duration of epilepsy preceding surgery [48,49,57,58]. In addition, patients with fixed acquired etiologies (eg, perinatal stroke) tend to have better motor and cognitive outcomes than those with progressive or congenital etiologies.
Corpus callosotomy — In a corpus callosotomy, the fibers of the corpus callosum are surgically divided. Typically, an anterior two-thirds callosotomy is performed first, with completion of the callosotomy done only if seizures persist after the first procedure [59]. Selective posterior callosotomy has also been explored [60].
Corpus callosotomy is primarily considered in patients with atonic seizures, which are most commonly seen in epileptic encephalopathies such as Lennox-Gastaut syndrome [61]. However, other reports suggest that a substantive reduction in generalized tonic-clonic seizures and other seizure types can result after corpus callosotomy, and that improvements in behavior, intelligence quotient (IQ), and overall quality of life are also possible [61,62].
Others — Stereotactic radiosurgery may be helpful for selected cases when the lesion is located where a conventional surgical approach is technically difficult or excessively risky [63]. More information is needed on long-term outcome before wider application of this procedure.
Psychosocial outcomes — The outcome of epilepsy surgery is typically measured by the degree of seizure control. Nonseizure outcomes, including psychosocial function and developmental outcomes, are less well-studied.
The available short-term data suggest that psychosocial outcomes at one year after epilepsy surgery are equivalent or better compared with ongoing medical therapy, especially among children with excellent postoperative seizure control [23,32,64-66]. Several observational studies with longer-term follow-up have suggested that IQ and other aspects of cognitive and psychosocial function improve over time compared with nonsurgical control patients, particularly in those patients who achieve seizure freedom or a reduction in antiseizure medications [23,67-70].
Antiseizure drug management after surgery — In some cases, epilepsy surgery affects a "cure," such that the patient is seizure-free off medication, but the risk for future seizure does not revert fully back to general population levels. Although seizure-free rates following temporal lobectomy may be as high as 70 to 80 percent, the longer patients are followed, the smaller this percentage becomes. A large study that examined seizure-free rates in over 600 adults 10 years after epilepsy surgery showed that only approximately 50 percent were still seizure-free [71]. Many patients therefore choose to remain on antiseizure medication following successful epilepsy surgery given their higher risk for seizure recurrence should they stop.
Observational studies on medication discontinuation after epilepsy surgery in children include the following series:
●The retrospective TimeToStop cohort study included 766 children who underwent epilepsy surgery over a seven-year period at one of 15 epilepsy centers across Europe and who began antiseizure medication reduction within a year of surgery [72]. The median times to antiseizure medication reduction and complete withdrawal were 13 and 29 months, respectively. Although earlier antiseizure medication withdrawal was associated with a higher risk for seizure recurrence, it did not affect long-term rates of seizure control or seizure freedom. This suggests that early antiseizure medication withdrawal may unmask incomplete surgical success sooner, identifying children who need ongoing antiseizure medication treatment and potentially sparing unnecessary continuation of antiseizure medications in others.
In a separate study that included 301 children from this cohort for whom preoperative and postoperative IQ measurements were available, antiseizure medication reduction was associated with improved postoperative IQ scores, independent of other variables [73]. The largest gains were achieved in those who were able to completely discontinue antiseizure medications (mean gain of 5.6 points).
●In one survey of 140 children, 102 had completely discontinued medication, and 90 percent were seizure-free off medication [74]. Among those whose seizures recurred, seven were able to regain seizure control with restarting medication, and three had recurring seizures that were not controlled by restarting medication.
●In another cohort of 80 children, 44 percent relapsed without medication [75]. Relapse rates were somewhat lower, 32 percent, in children with temporal lobe resection.
●Among 97 children undergoing temporal lobectomy, neocortical resection, or multilobar surgery, medication was discontinued in 68 and 57 remained seizure-free for two or more years [76]. Only 5 of the 97 patients continued to have seizures on medications.
Many centers recommend continuing medication for at least six months following surgery. If the child is seizure-free, then each drug is slowly tapered, one at a time, starting with the one that seemed least effective prior to surgery or the one thought to be causing the most adverse effects. Even if the child is not completely seizure-free, an attempt should be made to minimize the number of medications, hopefully to monotherapy.
VAGUS NERVE STIMULATION — Many children who fail antiseizure medications therapy and the ketogenic diet are not candidates for resective epilepsy surgery. Others may undergo resective surgery and still have seizures. Others may have seizures controlled but with significant, bothersome side effects. These patients may be candidates for vagus nerve stimulation (VNS). VNS is discussed separately. (See "Vagus nerve stimulation therapy for the treatment of epilepsy".)
THE KETOGENIC DIET — The ketogenic diet and related dietary modifications can be an effective treatment for patients with medically refractory epilepsy. This treatment, its indications, and potential adverse effects are discussed separately. (See "Ketogenic dietary therapies for the treatment of epilepsy".)
IMMUNOTHERAPIES — Some epileptic syndromes (eg, Rasmussen encephalitis and others) may have an immunopathogenesis and as such may be responsive to immunotherapies, such as intravenous immunoglobulin and corticosteroids. Observational data are limited and inconsistent for these therapies in syndromes such as Rasmussen encephalitis, Landau-Kleffner syndrome, and electrical status epilepticus in sleep [77-84]. Further controlled studies are needed in order to define which syndromes benefit from such treatment. In general, use of such therapies is restricted to the care of children with intractable epilepsy followed in tertiary-care, subspecialty epilepsy clinics. (See "Epilepsy syndromes in children", section on 'Developmental and epileptic encephalopathy with spike-wave activation in sleep (DEE-SWAS)'.)
The use of corticosteroid therapy in the treatment of infantile spasms and West syndrome is discussed in detail separately. (See "Infantile spasms: Management and prognosis".)
CANNABINOIDS — Safety and efficacy data are accumulating for cannabidiol (pharmaceutical), a component of cannabis (marijuana) most commonly investigated in medical studies [85-87]. Several randomized trials have demonstrated modest efficacy of a standardized preparation of cannabidiol oil for patients with Dravet syndrome (DS) or Lennox-Gastaut syndrome (LGS), and cannabidiol (pharmaceutical) is approved for the treatment of seizures associated with DS or LGS in patients ≥2 years of age. These data are reviewed separately. (See "Lennox-Gastaut syndrome" and "Dravet syndrome: Management and prognosis".)
However, for other indications, cannabidiol or other cannabis compounds should not be used outside of the context of a clinical trial. Similarly, self-treatment with smoked marijuana is not recommended [88].
Adverse effects of cannabidiol include diarrhea, fatigue, drowsiness, and changes in appetite [85-87].
Cannabidiol is a potent inhibitor of the CYP3A and CYP2C enzymes [89,90], which are responsible for metabolism of clobazam and many other antiseizure medications, including phenobarbital, phenytoin, carbamazepine, tiagabine, and valproate (table 3). In an open-label, prospective study of oral cannabidiol for drug-resistant epilepsy. a post hoc analysis found that patients on clobazam had a significantly higher responder rate (ie, a ≥50 percent reduction in motor seizures) compared with patients not taking clobazam (51 versus 27 percent) [91]. While metabolite levels were not followed in this study, a separate study found that clobazam metabolite levels rose by 60 percent with concomitant cannabidiol [92]. These effects may partially explain both toxicity and efficacy of cannabidiol in patients taking concurrent clobazam.
Importantly, long-term safety of cannabidiol has not been established, and significant concerns exist regarding the potential negative effects of chronic cannabis use on brain development, cognitive function, and school performance [93,94], particularly in children with drug-resistant epilepsy, who may have increased vulnerability to such effects.
Patients receiving cannabinoids should be monitored for adverse effects, including elevated liver function tests. Given the potential for drug-drug interactions with other antiseizure medications, closer serum monitoring of antiseizure medication levels may be prudent [95].
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 children".)
SUMMARY AND RECOMMENDATIONS
●When seizures do not respond as expected to initial antiseizure medication therapy, it is important to reconsider the seizure classification and the appropriateness of the antiseizure medication regimens that have been tried. In addition, clinicians should consider whether the diagnosis of epilepsy is accurate, as misdiagnosis is common. (See 'Evaluation' above.)
●The first antiseizure medication fails in 20 to 40 percent of children with epilepsy; lack of efficacy and side effects contribute roughly equally to treatment failure. Adding a second antiseizure medication is a reasonable next step when seizures are resistant to adequate doses of the initial drug. For children on polytherapy who continue to have inadequate seizure control, seizure frequency is sometimes improved when the number of antiseizure medications is reduced. (See 'Adding a second antiseizure medication' above and 'Inadequate response to multidrug therapy' above.)
●Transmucosal antiseizure medication administration (eg, rectal diazepam) is an effective and relatively safe emergency treatment for patients with prolonged seizures or seizure clusters in settings where intravenous access is not readily available. (See 'Home rescue therapy (transmucosal antiseizure medications)' above.)
●When a child has intractable epilepsy despite treatment with more than one antiseizure medication, referral to a comprehensive epilepsy center is appropriate to determine eligibility for epilepsy surgery and other treatment modalities, such as dietary therapy and vagus nerve stimulation (VNS).
•A number of effective surgical treatments are available (focal resections, lobar or multilobar resections, corpus callosotomy, hemispherectomy, and multiple subpial transection). The procedure selected depends upon the type and localization of the seizures. (See 'Epilepsy surgery' above.)
•The ketogenic diet can be an effective treatment for specific patients with medically refractory epilepsy. This treatment, its indications, and potential adverse effects are discussed separately. (See "Ketogenic dietary therapies for the treatment of epilepsy".)
●VNS is also effective in the treatment of refractory epilepsy when surgery is not an option. (See "Vagus nerve stimulation therapy for the treatment of epilepsy".)
