What is the optimal management of partial epilepsy uncontrolled by a first choice anticonvulsant?

What is the evidence of uncertainty?

A single small randomised study has compared the options of switching anticonvulsants or adding a second drug.8 No differences were detected, but the study was small and lacked sufficient power to detect clinically important differences.

A large number of placebo controlled regulatory studies in patients with refractory localised onset seizures have indicated some short term reduction in the frequency of seizures after adding a second anticonvulsant to the existing treatment.9 This potential benefit is, however, accompanied by an increased incidence of adverse events. The longer term effect of this risk-benefit ratio on patient perceived quality of life is uncertain. Studies in patients with less severe epilepsy (those in seizure remission on drug treatment and those with few or infrequent seizures at diagnosis) show that the benefits from improved seizure control are equally balanced by worse quality of life as a result of taking anticonvulsants.10 11

We also have no evidence on which second line drug is optimal. Over the past two decades, many newer anticonvulsants have been licensed for add-on treatment for localised onset seizures. Consequently, patients with refractory seizures have had their treatment regimen changed frequently, with uncertain benefits.

Is ongoing research likely to provide relevant evidence?

There are some existing and planned industry sponsored studies comparing the addition of different second line drugs. However, these are of short duration and are not suited to looking at important longer term patient based outcomes that are crucial to guide treatment choices in a chronic disorder such as epilepsy.

What should we do in the light of the uncertainty?

In patients with continuing localised onset seizures despite optimal first line therapy, consider adding a second anticonvulsant. A consensus evidence-free view is that when drugs are combined they should have differing primary mechanisms of action and no pharmacokinetic interaction.12 Most first line anticonvulsants (carbamazepine, lamotrigine, phenytoin, oxcarbazepine) act on sodium channels. Of the commonly used first line add-on anticonvulsants, our current preferred options include clobazam (which acts at GABA (γ-aminobutyric acid) receptors) or levetiracetam (which binds synaptic vesicle protein 2A).9 No evidence is available to support our choice of add-on—topiramate, pregabalin, and zonisamide are all reasonable alternatives.

A randomised controlled trial is urgently needed that compares adding placebo with adding one or two second line agents to existing optimally dosed treatment with carbamazepine, lamotrigine, phenytoin, or oxcarbazepine in patients with continuing localised onset seizures. Such a trial should evaluate longer term clinical, patient based, and health economic outcomes. Input from patients and carers is essential to help understand the balance between the effects of improved seizure control and the additional burden of drug treatment. In the meantime, patients should be offered the addition of a second line anticonvulsant, such as clobazam or levetiracetam, with appropriate counselling about risks and benefits and the opportunity to review outcomes over the short to medium term.