Efficacy of antiepileptic drugs in adults predicts efficacy in children: a systematic review

Abstract

Objective: Due to the challenges inherent in performing clinical trials in children, a systematic review of published clinical trials was performed to determine whether the efficacy of antiepileptic drugs (AEDs) in adults can be used to predict the efficacy of AEDs in the pediatric population.

Methods: Medline/PubMed, EMBASE, and Cochrane library searches (1970-January 2010) were conducted for clinical trials of partial-onset seizures (POS) and primary generalized tonic-clonic seizures (PGTCS) in adults and in children <2 and 2-18 years. Independent epidemiologists used standardized search and study evaluation criteria to select eligible trials. Forest plots were used to investigate the relative strength of placebo-subtracted effect measures.

Results: Among 30 adjunctive therapy POS trials in adults and children (2-18 years) that met evaluation criteria, effect measures were consistent between adults and children for gabapentin, lamotrigine, levetiracetam, oxcarbazepine, and topiramate. Placebo-subtracted median percent seizure reduction between baseline and treatment periods (ranging from 7.0% to 58.6% in adults and from 10.5% to 31.2% in children) was significant for 40/46 and 6/6 of the treatment groups studied. The ≥50% responder rate (ranging from 2.0% to 43.0% in adults and from 3.0% to 26.0% in children) was significant for 37/43 and 5/8 treatment groups. In children <2 years, an insufficient number of trials were eligible for analysis.

Conclusions: This systematic review supports the extrapolation of efficacy results in adults to predict a similar adjunctive treatment response in 2- to 18-year-old children with POS.

Figure 1. Algorithm for identification of clinical trials in adults and children 2–18 years of age

Figure 2. Efficacy comparison of differences in median % seizure reduction between baseline and treatment periods by drug for children and adults

A single row in the forest plot represents 1 trial or 1 trial regimen. In these plots, each effect measure is bounded by a confidence interval (CI). The null hypothesis for these analyses—that there would be no difference between drug-treated patients and those who received placebo—would yield a value of zero for each effect measure, indicated by the center line of the plot. An effect measure to the right of center favors drug treatment, while an effect measure to the left of center favors placebo. The CIs indicate the precision of that effect measure estimate. If the interval includes zero, this indicates that the p value was >0.05. inf ITT = inferential intention-to-treat; ITT = intention-to-treat; mITT = modified intention-to-treat; PP = per protocol.

Figure 3. Efficacy comparison of differences in ≥50% reduction in seizure frequency between baseline and treatment periods by drug for children and adults

A single row in the forest plot represents 1 trial or 1 trial regimen. In these plots, each effect measure is bounded by a confidence interval (CI). The null hypothesis for these analyses—that there would be no difference between drug-treated patients and those who received placebo—would yield a value of zero for each effect measure, indicated by the center line of the plot. An effect measure to the right of center favors drug treatment, while an effect measure to the left of center favors placebo. The CIs indicate the precision of that effect measure estimate. If the interval includes zero, this indicates that the p value was >0.05. inf ITT = inferential intention-to-treat; ITT = intention-to-treat; mITT = modified intention-to-treat; PP = per protocol.