New Anti-Epileptic Drugs Only
American Academy of Neurology (AAN) guidelines evaluating efficacy of new anti-epileptic drugs state only that lamotrigine is effective in children with newly diagnosed absence seizures (level B recommendation).
After ethosuximide was described as effective in absence epilepsy in 1958, a number of open-label, non-comparative studies where ethosuximide was added to existing treatment followed. However, at that time epilepsy syndromes were still poorly defined and trial methodology was not well developed. The overall seizure-free rate for these studies was 52%. There is one single-blinded, comparative, non-randomized study with placebo in untreated patients. One week of placebo preceded 8 weeks of treatment with ethosuximide. Fourteen of the 37 recruited patients had, in addition to clinically defined absence seizures, other seizure types. Absence seizures were controlled in seven patients. All these studies were poorly designed and they did not clearly show the percentage of patients with absence seizures which responded to ethosuximide.
Sodium valproate has been assessed in several open, non-comparative studies, in which the seizure-free rate has varied from 88% to 95%. Seven children with absence seizures were treated with valproate, and all but one became seizure free. Twenty-four-hour electroencephalography was performed before and during treatment. A dose-response relationship was suggested by the authors regarding reduction of spike-wave discharges, but the number of patients was too small to draw any conclusions.
Two single-blind, placebo-controlled trials with valproate have been published. Both included a period of placebo followed by a 10-week period of treatment with valproate. In the study by Villarreal et al., involving 25 adult patients with absence seizures (some with other seizure types and some receiving other anti-epileptic drugs), 10-week treatment with valproate resulted in 75% or greater reduction in spike and wave EEG discharges in 46% of patients. Nineteen patients had fewer absence seizures, which seemed to correlate with serum levels reaching 50-60 μg / ml. Another study, involving mainly children with refractory seizures, showed that 82% of children had at least a 75% reduction in the number of absence seizures. There was no correlation between plasma levels of valproate and response.
A double-blind study comparing ethosuximide with valproate looked at a treatment-naive group and a group of patients who were refractory to treatment with ethosuximide. The main outcome measure was the frequency and duration of 3-Hz spike and wave discharges with 12-h electroencephalography. A crossover design was used, in which valproate and ethosuximide were switched if the previously treatment-naive untreated group failed to achieve 100% control or the refractory group failed to achieve 80% of seizure control. There was a good response to treatment with either drug in the newly diagnosed group that was accompanied by reductions in spike and wave discharges. In the refractory group, 3 of 15 patients responded to valproate and 4 of the 14 patients responded to ethosuximide. There was no statistically significant difference in efficacy between ethosuximide and valproate or in reduction of spike and wave discharges. Pitfalls of these studies include the small number of patients and the fact that the design was not a true crossover, as patients who responded to the initial drug were not crossed over.
An add-on, comparative study of ethosuximide and valproate in 35 patients with mainly absence seizures found no significant differences between the drugs. By counting attacks (without EEG monitoring), they found improvement in 81% of patients in the valproate group and 71% of patients in the ethosuximide group. Another comparative, randomized, but not blinded, study found no significant difference in efficacy between the drugs. Increased efficacy with the combination of ethosuximide and valproate combination treatment was claimed in a small open-label study where neither drug alone proved effective. This was thought to be due partly to a pharmacokinetic interaction, because valproate inhibits the metabolism of ethosuximide, increasing the serum concentrations.
Common clinical practice is to choose valproate in patients with both absence and generalized tonic-clonic () seizures. And because 50% of patients with childhood absence epilepsy will develop generalized-onset tonic–clonic seizures, it can be argued that valproate is preferable because of its broader spectrum.
An ongoing study (National Institutes of Health childhood absence epilepsy trial) is analyzing the comparative efficacy and tolerability of valproate, lamotrigine and ethosuximide in patients with newly diagnosed absence seizures.
The benzodiazepine that has been more studied in the treatment of absence epilepsy is clonazepam. CZP was found to be superior to placebo in a single-blind, crossover study with 10 refractory patients. Eight of 10 patients became seizure free.
One double-blind, placebo-controlled study of patients (aged 3-15) was not a pure initial monotherapy trial, but rather a conditional, randomized conversion to placebo double-blind trial. In the study, 45 patients entered an open-label dose-escalation phase of lamotrigine lasting 5-25 weeks followed by a 4-week double-blind placebo-controlled phase in which patients with well-controlled absence seizures were randomized either to continue lamotrigine at their current dose or to be weaned to placebo. Overall, 28 patients were randomized to lamotrigine (n = 14) or placebo (n = 14). The proportion of patients remaining seizure free during the double-blind treatment phase was greater for lamotrigine (61%) compared with placebo (21%) (P = 0.03). The AAN rated the evidence provided by this study as class II, due to the enriched design. ILAE guidelines consider this class III evidence due to the design and short duration of the double-blind phase.
One study of lamotrigine as first-line therapy included 20 children with childhood absence seizures. After a mean follow-up of 10.8 months, 11 of 20 were seizure free, with another four having a greater than 50% decrease in seizure frequency. lamotrigine and valproate were compared in an open-label study involving 38 patients who were randomized to either valproate (n = 19) or lamotrigine (n = 19) and followed for 1 year. At the end of 12 months, no statistical difference in seizure-free rates was found between the two groups, although valproate acted faster in achieving seizure freedom.
One meta-analysis examined anti-epileptic drug efficacy and effectiveness for children with absence seizures. This compared ethosuximide, valproate and lamotrigine with a focus on four end-points: proportion of seizure-free children at 1,6 and 18 months after randomization; children with > 50% reduction in seizure frequency; normalization of the EEG; and adverse effects.
The majority of data used in this meta-analysis were from class III studies. The meta-analysis found insufficient evidence to inform clinical practice.
A small open-label trial of topiramate in five children with typical absence (untreated or unsuccessfully treated) reported one previously untreated child becoming seizure free, two improved at low dose (6 mg / kg / day) but not at higher doses, and two children (one previously untreated) showed no significant improvement.
In an open-label study, levetiracetam was administered to four patients with childhood absence epilepsy (as initial monotherapy in one and as add-on therapy in three). Three patients became seizure free and one patient had 50-75% seizure reduction. Clinical improvement was accompanied by a reduction or disappearance of generalized spike and wave discharges.
Because of its multiple mechanisms of action, zonisamide is thought to be a potentially broad-spectrum agent with efficacy for seizure types other than partial seizures. A recent retrospective, chart review study examined 15 patients with absence seizures, most of whom (89%) had received previous anti-epileptic drug therapy. Of these patients, 51% reported seizure freedom with zonisamide. This study, however, does not provide details as to whether the absences were typical or atypical, and no syndromic classification is provided. Several abstracts provide some suggestion that zonisamide may have some efficacy for absence seizures. In summary, there are no well-controlled studies and many of the published reports are preliminary reports of small numbers of patients.
Common Clinical Practice
First-line drugs to treat absence epilepsy are ethosuximide, valproate and lamotrigine, alone or in combination. Maximum tolerated doses must be reached before switching to a second monotherapy. When choosing the first drug, it is important to consider if the patient has only absence seizures or has other generalized seizure types in addition. Sodium valproate controls absences in 75% of patients and also has the advantage of controlling generalized tonic-clonic seizures (70%) and myoclonic jerks (75%). lamotrigine may control absences in possibly 50-60%, but may worsen myoclonic jerks. ethosuximide controls 70% of absences but it is undesirable as monotherapy if the patient has other types of generalized seizures. Small doses of lamotrigine added to valproate may have a dramatic beneficial effect. clonazepam and clobazam are second-line drugs for absence epilepsy; their use is limited by their adverse effects (mainly somnolence) and the development of tolerance in a significant proportion of patients. clonazepam may be useful as adjunctive drug especially in absences with myoclonic components. Carbamazepine (), vigabatrin (), gabapentin () and tiagabine () are contraindicated because they may exacerbate absence seizures. Phenytoin () and phenobarbital () are contraindicated because of their usual inefficacy. The place of topiramate, levetiracetam and zonisamide is still unknown. Gradual withdrawal of medication is recommended in patients with absence seizures who are seizure free for 1-2 years and have a normalized EEG. EEG confirmation of the seizure-free state is needed during this withdrawal period.