Abstracts

Over the past two decades, the efficacy endpoint in clinical trials of new AEDs as adjunctive therapy has been reduction in baseline seizure frequency, requiring [ge]4-wk baselines and up to 12-wk treatment periods. Time to n-th seizure (T[sub]n[/sub]) has been proposed as an alternative endpoint that could reduce study duration, eliminate the baseline period, and broaden inclusion criteria, thereby speeding enrollment. Using data from TPM clinical trials, we compared T[sub]n[/sub][sub] [/sub]and median % reduction in baseline seizure frequency for their abilities to detect treatment effects., Two methods were used to compare endpoints: 1) reanalysis of TPM clinical trial data using T[sub]n[/sub] as efficacy endpoint; 2) a simulation study with a model based on TPM data. Data were from six randomized, double-blind, placebo-controlled trials with topiramate as adjunctive therapy (200-1000 mg/day) in refractory partial-onset seizures. T[sub]n[/sub] for 3, 6, 9, and 12 seizures (T[sub]3[/sub], T[sub]6[/sub], T[sub]9[/sub], T[sub]12[/sub]) were estimated using Kaplan-Meier analyses. Cox proportional hazard regression, stratified by baseline 28-day seizure frequency ([lt]6, 6-22, [gt]22), was used to estimate the hazard ratio for each study arm (dose) vs placebo. Simulation study was performed as 12-wk dose-ranging trial with 4-wk baseline assessing 6 doses to yield treatment effects (difference from placebo in median % seizure reduction) of 0-50%., T[sub]n [/sub]analysis showed superiority of TPM over placebo, consistent with original analyses of median % seizure reduction. Simulation study results showed a linear relationship between T[sub]n[/sub] hazard ratio and median % seizure reduction. T[sub]n[/sub] analysis was generally less powerful than seizure rate analysis, although power increased with number of seizures (n[underline][gt][/underline]6) and with stratification according to baseline seizure frequency. Adequate power of the stratified T[sub]6[/sub] analysis for detecting clinically meaningful treatment effects (30% difference from placebo) could be attained with a reasonable sample size (80 patients/arm). A study with a prospective baseline and 12-wk follow-up would require 36 subjects to achieve the same power. In TPM clinical trials, 80% of patients experienced their 6^th seizure within 6 wks of study initiation., Stratified T[sub]n[/sub] analysis, with n=6, can provide adequate power to detect clinically relevant effect sizes of an AED as adjunctive therapy in a sample of 80 patients with refractory partial-onset seizures. The potential advantages of reducing clinical trial duration from 16 wks to 6 wks and potentially eliminating the prospective baseline period must be weighed against potential disadvantages such as need for larger sample sizes., (Supported by Johnson [amp] Johnson Pharmaceutical Research [amp] Development, L.L.C., Raritan, NJ.)