Clinical comparability of the new antiepileptic drugs in refractory partial epilepsy: a systematic review and meta-analysis

Purpose: Evaluate the clinical comparability of new antiepileptic drugs (AEDs) in partial refractory epilepsy. Methods: Systematic review of randomized trials (RCTs) comparing a new AED (add‐on treatment) with placebo or another AED. Primary outcomes: responder (≥50% seizure reduction) and withdrawal (tolerability) rates. Pooled estimates of odds ratios (ORs) and number needed treat/harm (NNT/NNH) taking into account baseline risk were derived by random‐effects meta‐analysis. Adjusted frequentist indirect comparisons between AEDs were estimated. Key Findings: Sixty‐two placebo‐controlled (12,902 patients) and eight head‐to‐head RCTs (1,370 patients) were included. Pooled ORs for responder and withdrawal rates (vs. placebo) were 3.00 [95% confidence interval (CI) 2.63–3.41] and 1.48 (1.30–1.68), respectively. Indirect comparisons of responder rate based on relative measurements of treatment effect (ORs) favored topiramate (1.52; 1.06–2.20) in comparison to all other AEDs, whereas gabapentin (0.67; 0.46–0.97) and lacosamide (0.66; 0.48–0.92) were less efficacious, without significant heterogeneity. When analyses were based on absolute estimates (NNTs), topiramate and levetiracetam were more efficacious, whereas gabapentin and tiagabine were less efficacious. Withdrawal rate was higher with oxcarbazepine (OR 1.60; 1.12–2.29) and topiramate (OR 1.68; 1.07–2.63), and lower with gabapentin (OR 0.65; 0.42–1.00) and levetiracetam (OR 0.62; 0.43–0.89). Significance: The differences found are of relatively small magnitude to allow a definitive conclusion about which new AED(s) has superior effectiveness. This uncertainty probably reflects the limitations of conclusions based on indirect evidence. The process of pharmacologic clinical decision making in partial refractory epilepsy probably depends more on other aspects, such as individual patient characteristics and pharmacoeconomics, than on available controlled randomized evidence.