Treatment of New Onset Epilepsy

Evidence-based Guidelines

The two major organizations for professionals with an interest in epilepsy have come to a consensus on the use of antiepileptic drugs (AEDs) for new onset epilepsy. Their report (French et al, Neurology 2004; 62: 1252-1260) describes how they reviewed reports of clinical trials of the new generation of AEDs to determine whether the results supported the use of that drug as first line treatment for patients. The method, called ”evidence-based,” uses established definitions to rate the “evidence” from clinical trials. The desirable “Level A” rating requires either one very-well designed study or two somewhat lesser studies, both with positive results demonstrating efficacy for the drug.

This type of analysis differs both from reports of the results of individual clinical trials and from regulatory approval by the Food and Drug Administration (FDA) to claim that the drug is effective for a specific diagnosis and type of patient. Individual clinical trials can show highly significant differences between treatment groups, but be biased by poor design, selection of comparator group or dose, post-hoc analyses (defined after seeing the main result), analyses of non-standard endpoints, or suffer from small numbers of patients starting or completing the study. Most clinical trials initiated by the pharmaceutical industry meet the criteria for good design and careful statistical analysis. They are designed specifically to meet regulatory standards set by the FDA in order to gain approval to market the drug. However, current FDA standards have been interpreted as requiring a placebo-treated group for comparison with the active drug. The reason for this standard is because no AED has been proven superior to placebo. It has been widely assumed, largely based on epidemiological studies, that patients receiving an AED have fewer seizures than untreated patients, but this has not been proven in a clinical trial. This results in a dilemma for doctors who will not put patients with partial-onset seizures at risk by leaving them untreated. The possibility of progression to a generalized tonic-clonic seizure that could result in injury or loss of a job is too great a burden for ethical consideration. However, the situation for mild seizure types is different. Parents often can bear the risk of waiting through a brief placebo-treatment period before initiating an AED because of concerns about the effect of drugs on their developing child. These children often have been experiencing absence or myoclonic seizures for months before the diagnosis. Therefore, a brief additional delay before treatment is ethically acceptable.

Interestingly, most AEDs widely used for initial treatment do not meet the current FDA criterion of proven superiority to placebo! Phenobarbital remains the most widely prescribed drug for seizures worldwide although no placebo-controlled clinical trial ever proved its efficacy. Similarly, phenytoin, has not been studied versus placebo. Carbamazepine probably is the most extensively-studied AED because it often is used as a comparator in studies of other AEDs where the study design is planned to demonstrate equivalence to carbamazepine. Nonethless, carbamazepine has not been compared to placebo to demonstrate efficacy. It has proven superior to another AED in only one trial (unpublished). These standard AEDs are allowed to indicate their FDA approval for initial treatment of epilepsy based on old rules (grandfathered into current practice). This positions drugs approved for initial treatment (for which no placebo-controlled studies have been performed) against newer AEDs that cannot achieve the approval status because such studies now are unfeasible (due to ethical constraints against using a placebo for partial-onset epilepsy). Essentially, doctors are in a Catch-22 situation.

Evidence-based guidelines are a way to address the FDA process which restricts what a pharmaceutical company can claim is an effective and safe use for a drug. These guidelines bring doctors out of the Catch-22 dilemma and into the clinical reality of using good evidence on which to base their selection of the most appropriate drug for each patient. With the publication of carefully developed guidelines, doctors now can feel comfortable prescribing any of the AEDs that meet Level A criteria, as needed.

The Paradox

The consensus on the use of AEDs for medically refractory epilepsy raises a paradox. All seven of the new AEDs approved are effective for use as adjunctive (add-on) therapy in patients with refractory partial or mixed seizure disorders. Yet only four drugs are effective for new onset (mostly non-refractory) epilepsy as a monotherapy. As the authors clearly state, the lack of recommendation for the other three drugs is based on a lack of evidence. None of other drugs were studied for monotherapy and found to be ineffective. Indeed, there has never been an AED studied that is effective as adjunctive therapy in refractory partial epilepsy that was not effective for monotherapy in new onset partial epilepsy. Common sense suggests that if a drug is helpful in the most difficult cases, it will likely be effective for the mildest cases. However, we cannot make recommendations without data from well designed clinical trial data. We are left with clinical judgment – an invaluable, but potentially dangerous guide. Thus, many epilepsy specialists pay careful attention to side effect profiles when selecting an AED for new onset epilepsy. Many epileptologists use drugs that are only indicated for adjunctive therapy as monotherapy. Whether this clinical practice is correct remains uncertain. We sorely lack head-to-head studies that compare the effectiveness or toxicity of new AEDs. However, most epileptologists favor newer AEDs to older and potentially toxic ones such as phenobarbital and phenytoin. In short, we need clinical trials that make direct comparisons among AEDs.

Back to top

Return to Articles and Publications