New antiepileptic drugs: a systematic review of their efficacy and tolerability

Abstract

Objectives : To evaluate the efficacy and tolerability of the newly developed antiepileptic drugs gabapentin, lamotrigine, tiagabine, topiramate, vigabatrin, and zonisamide in patients with refractory partial epilepsy. Design : Systematic review of published and unpublished randomised controlled trials of add-on treatment with new antiepileptic drugs. Subjects : 20 published and eight unpublished trials representing 3883 patients with refractory partial epilepsy. Main outcome measures : Proportion of patients who (a) showed 50% or greater reduction in frequency of seizures (50% responders) and (b) withdrew from each study for any reason. Results : Odds ratios (95% confidence intervals) relative to placebo for 50% responders were 2.29 (1.53 to 3.43) for gabapentin, 2.32 (1.47 to 3.68) for lamotrigine, 3.03 (2.01 to 4.58) for tiagabine, 4.22 (2.80 to 6.35) for topiramate, 3.68 (2.45 to 5.51) for vigabatrin, and 2.47 (1.36 to 4.47) for zonisamide. Odds ratios for withdrawal were 1.36 (0.75 to 2.49) for gabapentin, 1.19 (0.79 to 1.79) for lamotrigine, 1.81 (1.21 to 2.70) for tiagabine, 2.42 (1.43 to 4.11) for topiramate, 2.58 (1.26 to 5.27) for vigabatrin, and 5.70 (1.76 to 18.49) for zonisamide. Comparing results for each drug showed that all of the 95% confidence intervals overlapped, indicating that they were not significantly different in terms of efficacy and tolerability. Conclusions : All six drugs were significantly better than placebo at reducing frequency of seizures. These results do not allow an evidence based choice between these drugs as we have no conclusive indication of differences in efficacy or tolerability. Key messages: At present there is insufficient evidence to guide a choice between these new treatments We systematically reviewed randomised placebo controlled studies of supplementary gabapentin, lamotrigine, tiagabine, topiramate, vigabatrin, and zonisamide Each drug was significantly better than placebo at preventing seizures, but none was significantly different from the others in terms of efficacy or tolerability, though the confidence intervals were wide Randomised trials comparing active treatments are needed to further evaluate these drugs