Patterns of seizure frequency reduction in clinical trial participants with lower baseline seizure frequency

Inclusion and exclusion criteria of clinical trials for seizures aim to select representative participants with a high enough seizure frequency to evaluate the efficacy of treatment in a relatively short double-blind period. To inform the selection of seizure frequency-based inclusion criteria, we evaluated the association between baseline seizure frequency and reduction of seizure frequency in the double-blind period.

Using data from 11 double-blind placebo-controlled trials of antiseizure medications for either focal or generalized onset epilepsy, we evaluated the association of baseline seizure frequency with 50% responder rate and percent reduction of seizure frequency in maintenance. We identified four patterns based on the presence or absence of significant association (p < .05) in placebo, active treatment, both, and neither. We also evaluated whether the time to prerandomization monthly seizure count (T-PSC) design impacted these associations.

In 55% of trials (6/11), there was no significant association of maintenance seizure frequency change with baseline seizure frequency. In 19% of trials (2/11), there were parallel elevations in placebo and active treatment responses for lower baseline seizure frequency. In one trial (1/11), that shift was observed in placebo only, whereas there was a ceiling effect of high response in levetiracetam. In the remaining 19% of trials (2/11), there were more seizure frequency reductions in lower baseline seizure frequencies in active treatment but not placebo. These associations were not modified when the T-PSC design was used.

The association of the magnitude of change in seizure frequency with baseline seizure frequency was inconsistent across trials. In eight of 11 trials, these patterns did not reduce the magnitude of difference between active treatment and placebo and thereby may not reduce statistical power. In only one trial did elevated placebo response reduce the difference between active treatment and placebo. In two trials, active treatment appeared more efficacious in lower seizure frequencies.