Correct classification of the epilepsy type after the first generalised tonic-clonic (GTC) seizure is important for the choice of treatment. There is a high disease burden with uncontrolled GTC seizures, thus the goal of treatment is to reduce or prevent seizures. Clinical trials have shown that perampanel is effective for patients with refractory primary GTC seizures. In a real-world study of perampanel in a refractory epilepsy clinic population (FYDATA; Follow-up of 1 Year Data of paTients on perAmpanel), the overall retention rate at 1 year was 60%. There was additional efficacy in patients with secondarily generalised seizures and elderly patients. Slower titration than in the clinical trials can mitigate adverse events and improve tolerability.
Perampanel is a novel antiepileptic drug (AED) that targets the α-amino-3-hydroxy- 5-methyl-4-isoxazolepropionic acid (AMPA) receptor, a potentially critical pathway for the pathophysiology of epileptic seizures.
Burden and diagnostic challenges in GTC seizures
Concept of GTC seizures
Prof Eugen Trinka defined epilepsy according to John Hughlings Jackson’s observation that “A convulsion is but a symptom, and implies only that there is an occasional, an excessive, and a disorderly discharge of nerve tissue …”. Epilepsy is not a unified disease, having many causes that result in different seizure types. Therefore, antiepileptic drugs (AEDs) are licensed for specific seizure types rather than for the disease per se, and do not usually treat the underlying cause.
Seizures are classified as generalised, focal, or unknown. A focal seizure can evolve into a secondarily generalised tonic-clonic (SGTC) seizure or bilateral convulsive seizure. Generalised seizures originate at some point within, and rapidly engage, bilaterally distributed networks, including cortical and subcortical structures. Although individual seizure onsets can appear localised, the location and lateralization are not necessarily consistent from one seizure to another.1 Focal seizures originate within networks limited to one hemisphere, which may be discretely localised or more widely distributed. For each seizure type, onset is consistent from one seizure to another with preferential propagation patterns, which can involve the contralateral hemisphere.
The various types of generalised seizures are related to different degrees of excitatory (thalamoreticular system) and inhibitory (caudate nucleus) interaction in the cortex causing diverse symptoms. A generalised seizure can be conceptualised as a cortical lesion rapidly engaging bilateral networks with bilateral spread. The speed of this activity causes difficulty in distinguishing between primary generalised tonic-clonic (PGTC) and rapid SGTC seizures.
In clinical practice, the diagnostic process involves classification of seizure type and epilepsy syndrome, and assignment of the cause, which is necessary to select appropriate medication. The cause can be either an electroclinical syndrome or classified according to aetiology (genetic, structural, metabolic, immune, infectious, or unknown). While this concept appears to be straightforward, misdiagnosis after a first seizure occurs in 4.6–30% of patients,2,3
and can result in inappropriate treatment. Essential diagnostic procedures to distinguish focal from generalised seizures include neurological examination, electroencephalography within <24 hours of the first seizure, and magnetic resonance imaging.
Burden of GTC seizures
Epilepsy is associated with a high disease burden, with high rates of comorbidity,
|Morbidity||IRR (95% CI)|
|Emergency department visits||1.50 (1.49–1.52)|
|Hospital admissions||1.86 (1.84–1.88|
|Road traffic accident injuries||2.08 (1.81–2.39)|
|CI, confidence interval; IRR, incidence rate ratio.|
Table 1: Morbidity associated with non-adherence in epilepsy days [reprinted with permission from AAN Enterprises, Inc © 2008].6
mortality, psychological dysfunction, social stigmatisation, reduced quality of life, and decreased life expectancy. In Europe, mortality in patients with epilepsy is two- to three-fold that of the general population, with a clear correlation between seizures and standardised mortality ratio (SMR; seizure free, 1.4–1.6; persistent seizures, 2.4–3.3).4 In China, the SMR is 3.9, with proportional mortality rates of 30% each for injury and stroke.5
The mortality risk increases by up to three-fold when patients are not adherent to medication (hazard ratio, 3.32; 95% CI, 3.11–3.54).6 Non-adherence is also associated with significantly higher morbidity (Table 1). GTC seizures are also significantly correlated with injury (2.9-fold for severe injury [p=0.07] and 10.3-fold for mild injury [p=0.001]).7,8
There is an increased risk of sudden unexpected death in epilepsy (SUDEP) in patients who are not seizure free; 1–3 seizures per year increases the risk of SUDEP by 19-fold in patients with age at onset <16 years and 10-fold in patients with age at onset ≥16 years.9–11 Therefore, the main goal to prevent SUDEP is to reduce seizure frequency.
Correct classification of the epilepsy type after the first seizure is important for treatment. Diagnosis is by neurological examination, EEG <24 hours, and MRI. There is a high disease burden with uncontrolled GTC seizures, including increased morbidity and mortality. Measures for prevention and reduction of mortality include better treatments and increased adherence to therapy.
Current evidence and contemporary management of GTC seizures
Prof Terence O’Brien explained that management and treatment response of the different types of convulsive seizures varies, so it is important to make the correct diagnosis for seizure control. Regardless of the type of onset, all GTC seizures are disabling, and are a risk factor for injury, hospitalisation, and SUDEP.10–12 The risk of SUDEP increases with the number of GTC seizures, such that a patient with ≥3 seizures/year has >15 times the risk of SUDEP of a patient with no seizures.12
Most of the available AEDs are effective to some extent against focal-onset tonic- clonic seizures. However, only a limited number of AEDs are licensed for PGTC seizures, and 20–40% of patients with these seizures do not have good control. More robust clinical data are needed to guide the use of AEDs in patients with GTC seizures, but as GTC seizures tend to occur at a lower rate than focal seizures, study design and recruitment are challenging. Guidance on AEDs for newly diagnosed or resistant GTC seizures therefore comes from guidelines, evidence reviews, and expert opinion.
Randomised controlled trials
Only a few randomised controlled trials (RCTs) provide high-level evidence for treatment of patients with refractory PGTC seizures, and all are relatively small. RCTs of lamotrigine, levetiracetam, and topiramate versus placebo in refractory GTC seizures had similar results, with approximately 60–70% of patients having seizure reduction and ≥50% responder rates.13–17
The best evidence for newly diagnosed GTC seizures comes from the SANAD study, in which valproate prolonged time to treatment failure compared with lamotrigine or topiramate, and reduced the time to 12-month seizure freedom compared with lamotrigine.18 Thus, if there are no contraindications, valproate is the first-line AED for newly diagnosed PGTC seizures. If valproate is unsuitable, such as in women of childbearing potential, lamotrigine is
recommended as first-line treatment.19
Adjunctive therapy for refractory PGTC seizures includes clobazam, lamotrigine, levetiracetam, valproate, or topiramate. AEDs that may aggravate PGTC seizures include carbamazepine, gabapentin, oxcarbazepine, phenytoin, and vigabatrin.
Seizures and pregnancy
GTC seizures are particularly dangerous during pregnancy, with implications for both mother and child.19,20 GTC seizure frequency may increase during pregnancy in some women, possibly because of changes in AED pharmacokinetics and poor adherence to treatment due to concerns about the foetus. Data from the Australian Pregnancy Register compared the seizure control of pregnant patients treated with different monotherapies.21 Although valproate provided good seizure control compared with lamotrigine and topiramate,levetiracetam provides both low teratogenicity and better seizure control.
Perampanel in GTC seizures
Perampanel is the first non-competitive AMPA receptor antagonist. AMPA receptors are primary mediators of fast excitatory neurotransmission in the brain. It is thought that reducing fast-excitatory neurotransmission decreases seizure propagation.
Pooled analysis of three phase 3 trials of perampanel (studies 304, 305, and 306) examined the efficacy of perampanel in SGTC seizures.22 There was a dose effect, with perampanel 8 mg having the greatest efficacy at 62.9% reduction in seizures per 28 days.
French et al investigated the efficacy of perampanel in a phase 3 trial of PGTC seizures (study 332).23 The 50% responder rates were 64.2% for perampanel and 39.5% for placebo, and reductions in seizure frequency per 28 days were 76.5% and 38.4%, respectively (Figure 1). Among patients who completed the maintenance phase, seizure freedom was achieved by 36.8% and 13.0%, respectively. Perampanel resulted in slightly more treatment emergent adverse events (TEAEs) and TEAEs leading to dose adjustment than placebo, but there were no differences in severe and serious TEAEs. The most common AEs were dizziness, fatigue, somnolence, irritability, and weight gain.
In patients with bilateral convulsive seizures, it can be unclear whether the seizures are of focal or generalised onset. A pooled analysis of the four phase 3 trials included patients with either PGTC or focal-onset bilateral tonic-clonic seizures who had received perampanel 8 mg/day.24 Reduction in seizure frequency per 28 days was 65.5% for perampanel and 24.6% for placebo (p<0.0001). Even patients with severe epilepsy had greater reduction in seizure frequency per 28 days with perampanel (p≤0.01). The ≥50% responder rates were 61.8% and 37.8%, respectively, (p<0.0001), and seizure freedom was achieved by 26.9% and 12.6%, respectively.
GTC seizures are the most dangerous and disabling seizure type in terms of mortality, injury, and morbidity. Therefore, the goal of treatment is to reduce the number of or stop the seizures. However, there is a lack of high-level evidence to guide management, especially for newly diagnosed patients. Sodium valproate is considered the first-line AED for PGTC seizures, except in women of childbearing potential, for whom lamotrigine or levetiracetam can be considered. In patients with refractory PGTC seizures, RCTs show that perampanel, lamotrigine, levetiracetam, and topiramate are effective.
From clinical trials to clinical settings: Perampanel in real-world practice
Dr Vicente Villanueva described the primary phase 3 RCTs of perampanel, which involved more than 1000 patients with partial-onset seizures (POS) or idiopathic generalised epilepsy (IGE).23,25–27 However, extrapolation of RCTs into clinical practice is needed to include populations that are not enrolled in RCTs, such as elderly patients, paediatric patients, and those with comorbidities, and also to assess the long-term efficacy and safety.
To this end, several real-life studies of perampanel in Europe and Canada have been performed and published.28
The FYDATA (Follow-up of 1 Year Data of patients on perampanel) study was a multicentre, retrospective, 1-year, observational study that examined the efficacy and safety of adjunctive perampanel in patients with POS in Spain.29 A total of 464 patients aged 12 years or older were treated with perampanel as adjunctive therapy. Most patients had refractory epilepsy (mean number of prior AEDs, 7.8).
In the RCTs, the perampanel dosage was 2–12 mg for POS and 2–8 mg for IGE. Titration was 2 mg/week in both groups. In the FYDATA study, the median dose was 6 mg (range, 2–12 mg). In the real-world trials, the usual perampanel dose was ≤ 8mg (range, 2–12 mg), although this varied across studies, and may have been related to the size of the patients in different countries.
In the RCTs, the recommended titration was 2 mg/1–2 weeks depending on concomitant use of enzyme-inducing AEDs (EIAEDs). In the FYDATA study, the titration was slower at 2 mg/2–4 weeks, which can help to mitigate AEs. The retention rate was 60%, which was comparable with the other real-world studies with ~1 year follow-up.
In the RCTs, median seizure reductions were 36% for complex partial plus secondary generalised seizures (p<0.001) and 63% for secondary generalised seizures alone (p<0.001).22 In the real-world trials, the results for focal seizures were similar to those of the RCTs at 27–57%. One study reported 89% efficacy for secondarily generalised seizures alone.28 In the FYDATA study, the overall median seizure reduction over 1 year was 33% (last observation carried forward) and 58.3% in the completer group. In patients with secondarily generalised seizures, the median seizure reduction was 75% in the overall population and 100% in the completer group.
For patients with IGE in the RCTs, median seizure reduction was 76.5% (p<0.0001), which is similar to the result in the overall population in the FYDATA study. Thus, the RCT results in PGTC and SGTC seizures are reliable in a real-world population.
Other studies have also replicated these results.
When perampanel was given concomitantly with EIAEDs in the FYDATA study, the efficacy was reduced compared with concomitant non-EIAEDs. To reach the same level of response for concomitant EIAEDs, a higher perampanel dose was needed (Figure 2). Thus, it may be that a relationship exists between increases in perampanel plasma concentration and reductions in seizure frequency, and a slight increase in perampanel dose when given concomitantly with an EIAED might be required.
In the RCTs, treatment-related TEAEs occurring in 5% of patients included dizziness, fatigue, headache, somnolence, and irritability. In the FYDATA study, most AEs occurred within the first 6 months and were generally mild to moderate. Similar to the RCTs, the main AEs were dizziness and somnolence. Although irritability and behavioural AEs occurred more often than in the RCTs, patients with psychiatric comorbidity could be expected in the real-world trials. Patients with a psychiatric comorbidity were more likely to have psychiatric AEs, particularly those with personality disorder or hyperactivity compared with depression or anxiety. Slow titration (2 mg/3–4 weeks) can reduce the AE rate.
The results of a real-world study of perampanel in a refractory epilepsy clinic population were similar to the RCTs. The retention rate at 1 year was 60%, and there was additional efficacy in secondarily generalised seizures. There is no uniform policy about concomitant EIAEDs, but a slight dose increase may be required. Psychiatric comorbidity is associated with increased psychiatric AEs. Slower titration can improve the AE profile and bedtime intake can improve tolerability.
1. Berg AT, et al. Epilepsia 2010;51:676–85.
2. Stroink et al. Neurology 2003;60:979–82.
3. Uldall et al. Arch Dis Child 2006;91:219–21.
4. Trinka E, et al. Epilepsia 2013;54:495–501.
5. Ding D, et al. Lancet Neurol 2006;5:823–27.
6. Faught E, et al. Neurology 2008;71:1572–8.
7. Tomson T, et al. Epilepsy Res 2004;60:1–16.
8. Asadi-Pooya AA ,et al. Seizure 2012;21:165–8.
9. Tomson T, et al. Epilepsia 2005;46 (Suppl 11): 54–61.
10. Monte CP, et al. Seizure 2007;16:1–7.
11. Hesdorffer DC, et al. Epilepsia 2012;53: 249–52.
12. Hesdorffer DC, et al. Epilepsia 2011;52: 1150–9.
13. Biton V, et al. Neurology 2005;65:1737–43.
14. Biton V, et al. Epilepsy Behav 2010;19:352–8.
15. Trevathan E, et al. Pediatrics 2006;118:e371–8.
16. Berkovic SF, et al. Neurology 2007;69:1751–60.
17. Biton V, et al. Neurology 1999;52:1330–7.
18. Marson AG, et al. Lancet 2007;369:1016–26.
19. NICE clinical guideline 137. 2013.
20. SIGN national clinical guideline 70. 2003.
21. Vajda FJ, et al. Epilepsia 2014;55:1229–34.
22. Steinhoff BJ, et al. Epilepsia 2013;54:1481–9.
23. French J, et al. Neurology 2015;85:950–7.
24. O’Brien TJ, et al. American Epilepsy Society. 2015. Poster 2.250.
25. French JA, et al. Neurology 2012;79:589–96.
26. Krauss GL, et al. Neurology 2012;78: 1408–15.
27. French JA, et al. Epilepsia 2013;54:117–25.
28. Trinka E, et al. Acta Neurol Scand 2016;133:160–72.
29. Villanueva V, et al. Epilepsy Res 2016;126: 201–10.