Perampanel is a long-acting novel antiepileptic drug (AED) with a unique mechanism of action targeting the AMPA (α-amino-3-hydroxy-5-methyl-4-isoxazole propionic acid) receptor. Perampanel’s novel mechanism of action permits combination therapy regardless of the concomitant AEDs. Owing to its broad spectrum of activity, perampanel is suitable for both focal and generalised seizures. Its long elimination half-life allows once daily dosing. Up-titration can be done in 2 mg steps at 2- to 4-week intervals and perampanel dosing at bedtime mitigates some adverse events. A higher target dose may be needed with enzyme-inducing AEDs. Overall, perampanel is well tolerated and efficacious in drug-resistant focal and generalised epilepsies.
Perampanel is a novel α-amino-3-hydroxy- 5-methyl-4-isoxazole propionic acid (AMPA) receptor antagonist. In patients aged 12 years and older, perampanel is approved as adjunctive therapy for partial onset seizures (POS) with or without secondarily generalised seizures in Australia, Canada (in patients ≥18 years), the EU, Hong Kong, Iceland, Israel, Japan, Liechtenstein, Lebanon, Malaysia, Mexico, Norway, Philippines, Russia, Singapore, South Korea, Kuwait, Switzerland, Taiwan, Thailand, and the US, and for primary generalised tonic-clonic (PGTC) seizures in the US, EU, Canada (in patients ≥18 years), Hong Kong, Iceland, Japan, Liechtenstein, Norway, Philippines, Russia, South Korea, and Taiwan. Approval of perampanel for PGTC seizures is expected in Malaysia, Singapore, and Thailand.
Patient selection: Insights from real-world data
Prof Martin Brodie of the University of Glasgow, Glasgow, UK, described perampanel as a highly selective non-competitive AMPA-type glutamate receptor antagonist that has been available in the UK since 2012. As this is a novel drug, understanding of the pharmacology will enable clinicians to obtain the best outcomes for their patients. Absorption of perampanel is relatively slow at 30–120 minutes to peak concentration, which is the rationale for night-time administration.1 Oral bioavailability is excellent at 100%. The volume of distribution is 1.1 L/kg, and protein binding is approximately 95%. The elimination half-life is long at 105 hours (53–136 hours), which impacts the titration period; to achieve steady state, five half-lives are needed (i.e. approximately 525 hours or 22 days). Perampanel is metabolised largely by cytochrome P (CYP) 3A4 and CYP3A5, and clearance is increased by enzyme inducers.Thus, the titration time is impacted by concomitant administration of enzyme-inducing antiepileptic drugs (EIAEDs), particularly carbamazepine and phenytoin,which reduce the half-life of perampanel to approximately 30–50%.
Perampanel has few effects on other AEDs. Ketoconazole slightly inhibits perampanel metabolism prolonging its elimination half-life by 15%. High-dose perampanel (12 mg daily, but not 4 mg or 8 mg daily) causes a consistent decrease in levonorgestrel exposure.
Randomised controlled trials
The pivotal double-blind randomised controlled trials (RCTs) all had the same design. Patients with focal seizures who were receiving one to three AEDs were enrolled. The study populations included adolescents (older than 12 years) and adults in 40 countries. Patients were randomised to placebo or perampanel 2, 4, 8, or 12 mg. After a 6-week titration period, there was a 13-week maintenance period at the randomised dose. The total number of patients randomised was 1478.2
The greatest reduction in seizures (50% responder rate) was noted for secondarily generalised seizures (versus all partial seizures or complex partial plus secondarily generalised seizures) when compared with placebo (8 mg, p < 0.001; 12 mg, p < 0.01). There was no evident dose-response, but this is not usually seen with AEDs. As the patients were randomised to a fixed dose, the study design does not equate to realworld clinical practice. However, adverse event (AE) monitoring in the RCTs showed a consistent AE profile.
As perampanel is metabolised by CYP3A4 and CYP3A5, EIAEDs such as carbamazepine, phenytoin, phenobarbital, topiramate, oxcarbazepine, and eslicarbazepine should all reduce the metabolism of perampanel.3 Gidal et al found that concomitant EIAEDs (carbamazepine, oxcarbazepine, and phenytoin) reduced exposure to perampanel, but had no effect on the shape of the reduction in seizure frequency curve.4 Thus, halving the perampanel concentration by co-administration of EIAEDs reduces the effect of perampanel by half. This means that higher and faster titration of perampanel may be necessary to achieve the desired effect in the presence of concomitant EIAEDs. Interestingly, despite apparently inducing CYP3A4 and CYP3A5, phenobarbital does not appear to affect perampanel clearance.1 The reason for this is unclear, but overtreatment should be avoided when giving perampanel in conjunction with phenobarbital.
In the clinical trials of perampanel, in which patients were randomised to a fixed dose with a fast titration schedule, there was a dose-response relationship for the most common AEs of dizziness, somnolence, fatigue, irritability, nausea, and falls.
Patients with epilepsy may fall or become ataxic due to the condition and their drug load. Occasionally, falls may occur in the form of myoclonic jerks in response to perampanel.2 As this AE is dose related, reducing the perampanel dose while maintaining efficacy should reduce this risk.
In a study of perampanel 8 mg in PGTC seizures in idiopathic generalised epilepsy (IGE), the reduction in seizures per 28 days was 76.5% for perampanel versus 38.4% for placebo (p < 0.0001), 50% responder rate was 64.2% versus 39.5% (p=0.0019), and seizure freedom during the maintenance phase was 30.9% versus 12.3%.5 A total of 80.2% of patients were maintained at the target dose of 8 mg, confirming the efficacy of perampanel. However, perampanel was associated with more AEs of dizziness, fatigue, headache, somnolence, and irritability than placebo. Discontinuations occurred in 16% of patients taking perampanel versus 12% of those taking placebo.
Irritability and aggression are known AEs of perampanel, particularly at high doses. Additionally, many patients with refractory epilepsy have neuropsychological and psychiatric comorbidities, including depression, bipolar disorder, panic attacks, psychosis, attention-deficit hyperactivity disorder, and anxiety.6 Patients with these conditions are not usually included in RCTs, but when starting to treat these patients with perampanel in real-world clinical practice, unexpected psychiatric and behavioural AEs may become evident. In the phase 3 studies, psychiatric AEs and hostility and aggression mainly occurred in patients treated at the higher doses (8 or 12 mg).7 Thus, care is needed when titrating to the higher doses.
The major neurotransmitter systems implicated in the pathogenesis of aggression include serotonin, glutamate, noradrenaline, dopamine, and γ-aminobutyric acid, together with their respective receptors.8 Many of these mechanisms are also targeted in epilepsy pharmacology (Figure 1a).9 Similarly, several brain regions are implicated in both epilepsy and aggression (Figure 1b).9 Therefore, the effects of AEDs in patients who already have the potential to develop psychiatric and behavioural AEs may be underestimated. The terminology for these effects ranges in increasing severity from agitation, through irritation, impulsivity, anger, hostility, and aggression to violence. Both levetiracetam and perampanel should be avoided in patients with a history of aggression or violence.
In the pooled phase 3 studies, among 143 adolescents (age 12–17 years) receiving one or more doses of study drug (98 perampanel, 45 placebo), aggression was reported as an AE by eight patients (8%) in the perampanel group and none in the placebo group.10 Therefore, care is needed when prescribing perampanel to adolescent patients.
In a double-blind, placebo-controlled add-on study of 133 teenagers randomised 2:1 to perampanel (8–12 mg daily) or placebo, aggressive behaviour was reported in one patient (2.3%) in the placebo group and seven patients (8.2%) in the perampanel group.11 Four patients responded to dose reduction and no patients required perampanel withdrawal.
(b) brain regions implicated in both epilepsy and aggression [reprinted with permission from ASPET © 2016].9 AMPA, α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid; GABA, γ-aminobutyric acid; NMDA, N-methyl-d-aspartate; SV2A,synaptic vesicle glycoprotein 2A.
Real-world clinical study
A prospective audit was done of the first 54 patients with focal epilepsy to be treated with perampanel at the Western Infirmary in Glasgow.12 Patients had a wide range of monthly seizure frequency (median 4, range 1–60), were taking a median of two other AEDs (range 1–4), and had a median of three previously failed drug schedules (range 1–15). Three patients (6%) became seizure free for ≥6 months (one each at doses of 4, 8, and 12 mg), eight (15%) demonstrated ≥50% seizure reduction, and 17 (32%) had <50% seizure reduction. The median dosing in patients continuing treatment with adjunctive perampanel was 8 mg (range 4–14 mg). Twenty one patients (39%) withdrew because of AEs, which may have been partly due to fast titration. The commonest AEs were nausea, vomiting, ataxia, dizziness, and sedation, while two patients developed substantial weight gain and one had recurrent falls. Six patients (11%) developed neuropsychiatric complications (three developed depression and three had irritability and/or aggression). The median dose in patients who withdrew due to AEs was low (4 mg; range, 2–12 mg), as patients tended to know early in the schedule whether or not they could tolerate perampanel.
Perampanel is a long-acting novel AED with a unique mechanism of action. Perampanel has a long elimination half-life (average 105 hours), allowing once daily dosing at night (Table 1).13 Concomitant EIAEDs may reduce its efficacy, necessitating higher doses than with non-EIAEDs. Behavioural AEs may occur at high doses, especially in patients with psychiatric comorbidities.
Table 1: Perampanel titration and prescribing plan.1,13
- The elimination half-life is 53–136 hours
- The time to steady-state is 11–28 days
- Induced patients take 2 weeks and uninduced patients take as long as 4 weeks to reach steady state, so slow titration, up to every 4 weeks, will benefit patients
- Start with perampanel 2 mg at night, increasing at a minimum of 2 weekly intervals by 2 mg increments up to 6 mg
- Titrate even more slowly above 6 mg as necessary
- Consider higher daily doses in patients taking concomitant EIAEDs such as carbamazepine
- Switch to twice daily administration if peak dose tolerability appears to be an emerging problem
- Monitor behavioural AEs, particularly in doses exceeding 6 mg daily, planning to reduce the dose if necessary
- Long elimination half-life = slow titration
Evidence and use of perampanel as an early adjunctive therapy in epilepsy
Dr Vicente Villanueva of the Neurology Service, Hospital La Fe, Valencia, Spain, discussed the rationale for use of perampanel as early adjunctive therapy in clinical practice. Currently, the available AEDs include those that act on sodium channels, calcium channels, γ-aminobutyric acid, synaptic vesicle glycoprotein 2A, potassium channels, and glutamate, including AMPA, N-methyl-D-aspartic acid, and kainate.14–19 The mechanism of action of perampanel is unique, as it is the only drug to act as a non-competitive AMPA receptor antagonist in the glutamate pathway.
Mechanism of action
The aim of rational polytherapy is to increase efficacy (synergism) and decrease AEs (antagonism) by combining drugs on the basis of their mechanisms of action, and should include combinations for which there is evidence to suggest a favourable interaction. Perampanel may offer the advantage of new combinations, as it has shown efficacy regardless of the mechanism of action of the concomitant AEDs (including carbamazepine, valproic acid, lamotrigine, levetiracetam, and oxcarbazine), both in clinical trials and in realworld studies.5,20
The FYDATA (Follow-up of 1 Year Data of paTients on perAmpanel) study investigated the efficacy and safety of adjunctive perampanel in patients with POS in real-world clinical practice in Spain.21 This was a multicentre, retrospective, 1-year, observational study in patients aged 12 years and older who had at least one POS in the previous year. Treatment with perampanel was given according to the usual clinical practice at the centres. Most of the patients had complicated treatment histories, having tried a mean of 7.8 prior AEDs.
Comparison of perampanel in patients taking sodium channel blockers with those who were not taking sodium channel blockers found no statistical difference in ≥50% responder rate (26.2% vs 36.7%; p=0.270). Owing to the novel mechanism of action of perampanel it is possible that it could have an additive effect to many other AED classes.
Spectrum of activity
The spectrum of activity may be one of the most important advantages of an AED. When first diagnosing epilepsy, the type of seizure may be unclassified in approximately 20% of patients after 1 seizure and 12% after more than one seizure.22 Undetermined seizure type at the time of diagnosis makes it difficult to select an appropriate broad-spectrum AED that will not aggravate seizures.23,24 The currently available broad-spectrum first-line options include valproate, lamotrigine, topiramate, levetiracetam, or zonisamide.25–29
In 2014, the European Medicines Agency strengthened warnings on the use of valproate in women and girls due to the risk of malformations and developmental problems in babies who are exposed to valproate in the womb.30 Preclusion of valproate as a treatment option has increased the difficulties of treating women and girls with generalised or unclassified epilepsy. However, perampanel has provided another treatment option, as it has been shown to be effective in IGE. Perampanel reduced PGTC seizure frequency by 76.5% (vs 38.4% for placebo; p < 0.0001) and achieved a 50% PGTC seizure responder rate of 64.2% (vs 39.5% for placebo; p=0.0019),31 which is similar to the result for levetiracetam for PGTC seizure freedom.32 Similarly, there is additional efficacy of perampanel in secondarily generalised tonic-clonic seizures over focal seizures and compared with some other AEDs.33,34
Extrapolation of this result to realworld clinical practice is shown in the FYDATA study. In patients with secondarily generalised seizures and focal seizures, median seizure reduction was 33.0% in the overall population (last observation carried forward [LOCF]) and 58.3% in patients who completed 1 year (Figure 2a).
For secondarily generalised seizures, the median seizure reduction was 75.0% in the overall population (LOCF) and 100.0% in patients who completed 1 year (Figure 2b). In the per protocol analysis, 7.2% of patients were seizure free, 26.8% had >50% response, and 43.6% had some improvement and, in the completer analysis, 10.3% were seizure free, 40.2% had >50% response, and 62.3% had some improvement.
This effect has been reproduced in other real-world clinical studies. In a multicentre survey of perampanel in Germany and Austria, 57% of patients with secondarily generalised seizures responded to treatment and 32% were seizure free after 6 months.35 In a UK audit of perampanel, 57.5% of patients with generalised tonic- clonic seizures achieved ≥50% seizure reduction. Of eight patients with IGE in this group, six showed seizure reduction.36 Perampanel has also shown efficacy for intractable seizures in patients with Lafora disease37 and Lennox-Gastaut syndrome with related encephalopathies.38 In the patients with Lennox-Gastaut syndrome, there were some behavioural disturbances, but also some improvement in alertness.
In the FYDATA study, adjunctive perampanel resulted in better efficacy in patients who had early add-on therapy (≤5 prior AEDs) than in those given perampanel at a later treatment stage. Thus, early treatment with perampanel may have better results. Adherence is a problem for patients with epilepsy, who may forget to take their medication and have breakthrough seizures, and clinicians may underestimate or overestimate the effect of a drug resulting in AEs or polypharmacy, decreasing a patient’s quality of life.39 Adherence is improved with a reduced number of intakes.40 Electronic monitoring studies have shown that increasing frequency of drug doses per day is associated with decreasing adherence, with a once-daily dose showing the best adherence.41
Elderly people often have polypharmacy, which is a cause of non-adherence. In the FYDATA study, the elderly population had a good response to perampanel, with fewer AEs than the younger population. This was also found in the Salzburg audit, which resulted in a high retention rate and lower AE profile among the elderly population.20 Additionally, behavioural disturbance was less common among the elderly population.
Quality of life of patients with epilepsy can be affected by seizures, psychiatric comorbidity, and cognition. Cognitive issues can influence QoL scores by up to 10%.21 A study of the cognitive effects of perampanel in teenagers (age 12–18 years) found that most of the results were similar to those of placebo.11 Speed of memory and continuity of attention were worse with perampanel, but quality of episodic memory was improved with perampanel.
In the FYDATA study, only 3% of patients had memory disturbance. Hyperactivity can be viewed as a behavioural problem, but it may also be due to greater alertness. However, clinicians can learn how to manage these effects and make the best use of perampanel.
In the UK audit, there were few effects on memory or confusion.36 In the Glasgow
prospective audit, very few patients withdrew because of cognitive problems,12 and no patients had cognitive issues in the multicentre survey from Germany and Austria.35
In the FYDATA study, AEs were not dose related, but were more related to titration. When up-titration was delayed from 2 mg every 2 weeks to 2 mg every 3–4 weeks, AEs (including psychiatric AEs) were reduced. The reason for this is the long half-life of perampanel. The UK audit also found that slower titration resulted in fewer AEs.36
In real-world clinical practice, the starting dose of perampanel is 2 mg for at least 2 weeks, which is increased to 4 mg for the first appointment. However, individual
patient characteristics such as size and weight, and presence or absence of EIAEDs
may help in deciding the optimal start dose and target dose.
Perampanel’s novel mechanism of action permits combination therapy regardless of the concomitant AEDs. Perampanel’s broad spectrum of activity is suitable for focal and generalised seizures, with better results if the number of prior AEDs is low. Once-daily dosing can improve adherence, particularly for elderly patients. Cognitive AEs are low in both randomised controlled trials and real-world studies, and can be minimised with slow titration and night-time administration.
Strategies to monitor and mitigate common AEs with perampanel
Prof Eugen Trinka of the Department of Neurology, Paracelsus Medical University,
Salzburg, Austria, explained that the AEs of perampanel can be mitigated as clinicians
learn how best to use the drug. From the real-world evidence presented to date, clinicians throughout Europe are using similar prescribing methods to reduce AEs in their clinical practices.
Common AEs of perampanel: study evidence
Safety data from the phase 3 studies (304, 305, and 306) show that the most frequent AEs occur at the higher perampanel doses of 8 and 12 mg, and include dizziness, somnolence, and headache.33 The incidence of headache with perampanel 4 and 8 mg was similar to that for placebo (all 11%)and that for perampanel 12 mg was 13%. Dizziness (32% and 43% for perampanel 8 and 12 mg, respectively) and somnolence (16% and 18% for perampanel 8 and 12 mg, respectively) occurred more frequently, but were mainly mild to moderate in intensity. Patients tended to experience dizziness immediately after drug administration during the course of the evening.
For the time of onset and duration of the most common AEs, dizziness, somnolence, and fatigue all disappeared over time. The median time of persistence of dizziness was 50 days. Although this may seem to be a fairly long time to experience dizziness, the duration is limited. Hostility/aggression also occurred in a dose-dependent manner (12% and 20% for perampanel 8 and 12 mg, respectively). A fixed titration schedule was used in the RCTs, but slower titration could reduce the occurrence of AEs in the real world.
Common AEs of perampanel: personal experience
A real-world retrospective analysis of the treatment response and AEs of add-on therapy with perampanel was done at the Paracelsus Medical University, Salzburg, Austria from September 2012 to January 2015.43 The study enrolled 125 eligible patients. The median duration of epilepsy was 17 years (range, 0.01–61.4 years) and the median remission was 3 months (range, 0.00–120.0 months). Seizure freedom was considered statistically significant at three times the longest seizure-free interval in an individual patient. Two-thirds of patients (66%) had symptomatic epilepsy, one-third (31%) had epilepsy of unknown cause, and three (3%) had IGE.
Patients were taking a broad range of baseline medications, including phenytoin,
valproate, levetiracetam, lamotrigine, lacosamide, and carbamazepine. A total of
44 patients (35%) were taking concomitant EIAEDs (carbamazepine, oxcarbazepine, phenytoin, and topiramate), and these patients were also analysed separately. The median number of previous AEDs was 4 (range 1–14) and the median number of concomitant AEDs was 2 (range, 1–4). Median seizure frequencies per month were 3.0 (range, 0.1–90.0) for simple focal and complex focal seizures and 0.3 (range, 0.0–45.0) for secondarily generalised tonic- clonic seizures.
Overall, the retention rate was 59%. When patients tolerate perampanel and respond, there is a high rate of seizure freedom, both with and without concomitant EIAEDs. However, there was a significant difference in seizure freedom between patients taking and not taking concomitant EIAEDs (n=5 and n=28, respectively; p=0.002). If patients do not tolerate the drug or do not respond at an early stage, it is not necessary to continue the treatment. The 50% responder rate was 35% and the rate of seizure freedom at 3 months was 27%.
The number of AEs increased with higher perampanel doses, up to 8 mg (only three patients were taking 10 or 12 mg). There was no difference in AEs between patients taking and not taking concomitant EIAEDs. At a median dose of 8 mg, 32% of patients had dizziness, again with no difference between the EI and non-EI groups. Overall, 33% of patients had dizziness, 12% had fatigue, 4% had nausea, 6% had cognitive deficit, 5% had speech problems, 4% had gait disorder (ataxia), and 7% had psychiatric AEs (abnormal thinking).
Fatigue was reported in 12% of patients, but the term is difficult to define, and correlates with depressive mood and sleep problems.44 Sleep problem is an active symptom of depression. Thus, some patients who rate fatigue as an AE may have depression, and could benefit from antidepressant therapy.
In total, 18 patients (13%) had dose reduction due to AEs, and there was no difference between the EI and non-EI groups. Seventeen patients (94%) had dose reduction of 2 mg with improvement in AEs. Overall, 51 patients (41%) stopped taking perampanel. Twenty five patients (20%) withdrew due to AEs. Other reasons for withdrawal included no improvement, reimbursement issues, seizure freedom, and unknown reasons.
Perampanel improved seizure control in this group of patients with difficult to treat focal epilepsy. The responder rate was 35%, and 6% of patients were seizure free for more than three times the longest remission. The median dose was 6 mg. AEs were usually mild, and dizziness was the most frequent AE; 94% of patients had improvement of AEs after dose reduction. Therefore, starting treatment with a low dose, and up-titrating slowly can reduce the AEs. Administration immediately before bedtime reduces AEs and increases tolerability. There was no significant difference in AEs between the EI and non-EI groups, but seizure control was better if no concomitant EIAEDs were taken (p=0.002). The slightly lower seizure control with concomitant EIAEDs can be managed by titrating to a slightly higher target dose.
Perampanel in elderly people
Interestingly, there was a better response rate (35% seizure free) and good tolerability in elderly patients (older than 60 years) than in younger patients. Elderly patients might be expected to have more vertigo, but the rate for patients older than 60 years was 15.0% versus 40.0% for adults younger than 60 years. Elderly patients also had less nausea (0.0% versus 7.7%), but did have more fatigue (20.0% versus 4.6%).45
Overall, perampanel is well tolerated and efficacious in drug-resistant focal and generalised epilepsies. Perampanel is given once daily, with up-titration in 2 mg steps at 2- to 4-week intervals. The common AEs observed with perampanel in clinical practice, including vertigo/dizziness, gait disturbance, and fatigue, can be reduced by a slow titration rate and by dose reduction (in 2 mg increments). A higher target dose may be needed with EIAEDs. Communication of information about possible AEs aids compliance, and taking perampanel immediately before sleep mitigates AEs such as dizziness. For AEs of aggression/hostility and irritability, a patient’s previous history, co-medication, and age should be considered. The best motivation is seizure control, and some AEs can be tolerated if a patient has a good response.
1. Brodie MJ. Antiepileptic Drugs. 4th Ed. 2015.
2. Steinhoff BJ, et al. Epilepsia. 2013;54:1481–9.
3. Brodie MJ, et al. Epilepsia. 2013;54:11–27.
4. Gidal BE, et al. Epilepsia. 2013;54:1490–7.
5. French JA, et al. Neurology. 2015;85:950–7.
6. Lin JJ, et al. Lancet. 2012;380:1180–92.
7. Ettinger AB, et al. Epilepsia. 2015;56:1252–67.
8. Comai S, et al. J Clin Psychopharmacol. 2012; 32: 83–94.
9. Brodie MJ, et al. Pharmacol Rev. 2016;68: 563–602.
10. Rosenfeld W, et al. Eur J Paed Neurol. 2015;19:435–45.
11. Meador KJ, et al. Epilepsia. 2016;57:243–51.
12. Brodie MJ, et al. Epilepsy Behav. 2016;54: 100–3.
13. Brodie MJ. CNS Drugs. 2015;29:893–904.
14. Lason W, et al. Pharmacol Rep. 2011;63: 271–92.
15. White HS, et al. Intern Revi Neurobiol. 2007;81:85–110.
16. Levy RH, et al. Antiepileptic Drugs. 5th ed. 2002.
17. Ceolin L, et al. Neurochem Int. 2012;61: 517–22.
18. Kellinghaus C. Ther Clin Risk Manag. 2009;5:757–66.
19. Bonifácio MJ. European Congress on Epilepsy. 2012. Abstr p422.
20. Trinka E, et al. Acta Neurol Scand. 2016;133:160–72.
21. Villanueva V, et al. Epilepsy Res 2016;126: 201–10.
22. Tchalla AE, et al. Epilepsia 2011;52:2203–8.
23. Sazgar M, et al. Pediatr Neurol. 2005;33: 227–34.
24. Gelisse P, et al. Epilepsia. 2004;45:1282–6.
25. Glauser T, et al. Epilepsia. 2006;47:1094–120.
26. French JA, et al. Neurology. 2004;62:1252–60.
27. Bergey GK. Epilepsia. 2005;46(Suppl 9): 161–8.
28. Scottish Intercollegiate Guidelines Network. A national clinical guide. April 2003.
29. National Institute for Clinical Excellence. Technology appraisal 76. March 2004.
30. European Medicines Agency. EMA/709243/2014. 21 November 2014.
31. French J, et al, Neurology. 2015:85:1–8.
32. Berkovic SF, et al. Neurology. 2007;69: 1751–60.
33. Steinhoff B, et al. Epilepsia. 2013;54:1481–9.
34. Hemery C, et al. Epilepsia. 2014;55:1789–99.
35. Steinhoff BJ, et al. Epilepsy Res. 2014;108:986–8.
36. Shah E, et al. Seizure. 2016;34:1–5.
37. Eren F, et al. International Epilepsy Congress. 2015. Poster 0208.
38. Genton P, et al. International Epilepsy Congress. 2015. Poster 0204.
39. Modi AC, et al. Epilepsy Behav. 2012;25:481–4.
40. Claxton AJ, et al. Clin Ther. 2001;23:1296–310.
41. Saini S, et al. Am J Manag Care. 2009;15: e22–e33.
42. Giovagnoli AR, et al, Epilepsy Res. 2014;108:1461–8.
43. Rohracher A, et al. Ther Adv Neurol Dis 2016. doi: 10.1177/1756285616661115.
44. Leite Neves GSM, et al. Arquivos de Neuro-Psiquiatria 2013;71:533–6.
45. Granbichler C, et al. American Epilepsy Society. 2014. Abstract 2.068.