New Antiepileptic Drugs: A Bit of Bad News

The new generation of AEDs is not perfect. Is anyone surprised by this?

In 1993, the release of felbamate (Felbatol) heralded the wave of new antiepileptic drugs (AEDs). These agents variously possessed characteristics such as fewer cognitive adverse effects and fewer drug interactions compared with the older agents. In general, the new generation of agents bypassed such problematic adverse effects as the nonlinearity seen in phenytoin use and the severe teratogenicity of valproic acid. However, as expected, these newer drugs are not perfect, nor are they panaceas. Recently, illustrations of this were published.

Glauser et al. just published the results of a large NIH sponsored trial of ethosuximide (ESM), valproic acid (VPA) and lamotrigine (LTG) for the treatment of childhood absence epilepsy. This was a multicenter double-blind randomized trial. 453 children between 2.5 and 13 years old were randomized. Drug effectiveness, a combined measure of efficacy and tolerability, was chosen as the determinant of benefit, as this is a "real world" measure of drug effect. The various components of this were combined into a treatment failure assessment, which was the primary outcome measure.

47% of children did not reach treatment failure (eg, had adequate seizure control without unacceptable adverse effects) by weeks 16 or 20. ESM (53%) and VPA (59%) treatments were better than LTG (29%) Odds ratios for not reaching treatment failure were 2.66 for ESM v. LTG and 3.34 for VPA v. LTG. This is a very robust difference. Reasons for treatment failure were about evenly divided between inadequate seizure control and adverse effects. Most patients with inadequately treated seizures were in the LTG cohort, and adverse effects were fairly evenly divided between treatment arms. Nausea, GI upset, somnolence and headache were rather more common in the ESM and/or VPA groups compared to the LTG group, but these adverse effects were said to be generally mild and transient. The authors report 13 rashes leading to treatment failure, but they do not report these data stratified by AED.

A cognitive function battery was completed in 316 patients. A specific cutoff score was used to determine relative test performance. The percentage of VPA patients (49%) compared to ESM patients (33%) reaching this cutoff was statistically significant (OR 1.95). VPA was also inferior to LTG (24%; OR 3.04). ESM and LTG did not separate. Also of note: Many more LTG patients (58.9%) were able to reach the maximal titration dose compared to VPA (20.5%) or ESM (17.5%).

So, LTG was overall less effective than ESM or VPA. The latter agents had somewhat more adverse effects, but they typically cleared over time. The authors conclude that the data imply that "ethosuximide is the optimal initial empirical monotherapy for childhood absence epilepsy."

The other trial, published in Neurology by Salinsky et al., examines the cognitive adverse effects of pregabalin (PGB) in healthy volunteers. One of the touted benefits of this agent is its freedom from such adverse effects, so this is a significant study. This was a double blind, placebo controlled, parallel group study of 12 weeks' duration. 32 patients were randomized, and 30 (94%) completed the study. PGB was titrated to 300 mg. b.i.d. on a flexible schedule. Multiple cognitive measures, as well as a subjective neurotoxicity scale, were obtained at a number of visits, but 6 cognitive tests were selected pre-hoc to minimize the statistical problems of analyzing a large number of scales.

The results showed problems with PGB. Both of the patients who dropped out were in the PGB group. Three of the 6 index tests showed significant differences compared with placebo at the week 12 testing, although all of the differences were graded as mild. In the overall set of tests, 5 of 20 measures favored placebo, and one favored PGB: The latter finding is well explained as a chance occurence. On the subjective toxicity measures, cognitive toxicity was significantly more common and of larger magnitude in the PGB group compared to placebo, but fatigue was not.

So, an older drug is more effective than a newer agent for childhood absence epilepsy. And, the absence of cognitive adverse effects with some of the newer drugs may be a bit overrated. Unfortunately, the new generation of AEDs is not perfect. Is anyone surprised by this?