Abstracts

Rationale: ESL is a novel once-daily anticonvulsant, extensively converted after oral administration to eslicarbazepine, which blocks voltage-gated sodium channels. In the absence of head-to-head trials among AEDs, network meta-analysis can generate evidence regarding their relative effectiveness as adjunctive therapy, and allows the calculation of the number of patients needed to treat in order to achieve a specified benefit. The objective of this study was to compare the relative effectiveness of effective doses of eslicarbazepine acetate (ESL) to that of recently-approved anti-epileptic drugs (AEDs) for adjunctive treatment of partial-onset seizures (POS) in adults. Methods: A systematic literature review of all AEDs approved by the FDA in the last 10 years for the adjunctive treatment of POS in adults was conducted, and all phase III placebo-controlled trials for ESL were identified. AEDs with a monotherapy indication or a non-POS indication, monotherapy trials, trials with known data quality issues, and ineffective doses (per approved label, or clinical trial results) were excluded. When publications for approved AEDs were not available, clinical trial results were consulted on the FDA website. The 50% responder rate (R50) during the maintenance period was extracted from the ITT population for the high- and low-dose formulations of each compound, and used to calculate the number needed to treat (NNT). The R50 for the double-blind period was used if the maintenance period R50 was not available. An indirect comparison was performed via Bayesian network meta-analysis. Results: Fourteen studies met inclusion/exclusion criteria (3 each for ESL, lacosamide (LAC), ezogabine (EZO), and perampanel (PMP); one each for oxcarbazepine extended-release (OXC) and levetiracetam extended-release (LEV)). The estimated R50 for the high-dose formulations were 41.8% for ESL 1200mg, 39.3% for EZO 1200mg, 38.8% for LAC 400mg, 37.1% for PMP 12mg, 33.9% for LEV 1000mg, and 32.6% for OXC 2400mg (see table). Among the low-dose formulations, the R50 were 35.2% for EZO 600mg, 33.2% for ESL 800mg, 32.3% for PMP 4mg, 31.6% for LAC 200mg, and 28.5% for OXC 1200mg. The NNTs for the high doses were: 4.9 for ESL 1200mg, 5.5 for EZO 1200mg, 5.7 for LAC 400mg, 6.3 for PMP 12mg, 7.9 for LEV 1000mg, and 8.8 for OXC 2400mg. The NNTs for the low doses were: 7.1 for EZO 600mg, 8.3 for ESL 800mg, 9.0 for PMP 4mg, 9.6 for LAC 200mg, and 13.7 for OXC 1200mg. Conclusions: This network meta-analysis found that higher medication doses were associated with increased efficacy, compared to lower doses. Compared to other AEDs recently approved for adjunctive treatment of partial-onset seizures, ESL 1200mg had the highest overall efficacy and lowest NNT among the high dose formulations, while ESL 800mg had the second highest overall efficacy, and the second-lowest NNT among the low-dose formulations. The reader should bear in mind that the validity of NMAs is limited compared to that of head-to-head RCTs.