Placebo responses in randomized trials of antiepileptic drugs.

This meta-analysis of published, randomized, controlled trials (RCTs) of antiepileptic drugs in adults with focal drug-resistant epilepsy was performed to estimate a mean placebo effect, to evaluate variability in placebo response rates, to investigate associations between placebo effect rates and study characteristics, and to determine whether there were changes in placebo response rates over recent years in RCTs (so-called "placebo drift"). One hundred ninety-eight potentially appropriate studies were identified after MEDLINE search and carefully reviewed. Twenty-seven RCTs (with 5662 randomized patients, including 1887 patients in placebo arms) were included in the meta-analysis. A random effects meta-analytic model estimated the pooled placebo response at 12.5% (95% CI: 10.03-14.94%). A statistically significant correlation between baseline median seizure frequency and placebo response rates was not observed. "Placebo drift" was not considered statistically significant.