Abstracts

This is a Late Breaking abstract

Rationale: A 2011 meta-analysis reported placebo response had been increasing in randomized controlled trials (RCTs) of anti-seizure medications for treatment- resistant epilepsy. Higher placebo rates reduce the efficacy of these randomized clinical trials (RCTs) making them more expensive. Nevertheless, there remains a strong need for new treatment options for patients with epilepsy. Hence, it is critical to find ways to reduce placebo responses in RCTs, making them easier and less expensive to conduct. We hypothesized that inclusion criteria may impact placebo responses._x000D_
Methods: We conducted a literature search with PubMed using the following search strings: epilepsy, drugs, randomized, placebo, double-blinded, “phase 3” or “phase III,” or “phase 2” or “phase II,” or “phase 2/3.” Articles published from 2000 to 2022 were included. The resulting studies were evaluated for relevance to outpatient epilepsy treatments and tabulated relevant data from each study including year of publication, number of patients in placebo group, inclusion criteria, 50% responder rate (RR50) and median percentage change (MPC) in seizure frequency in the placebo groups. We noted if inclusion criteria required a historical seizure frequency, measured baseline seizure frequency, or both and how it impacted the result. Correlations were assessed with Pearson correlation coefficient._x000D_
Results: Placebo responses have not been increasing between 2000 and 2022 (N=46, RR50 R²=0.00 and MPC R²=0.01). However, within the studies (N=11) that had both historical and baseline inclusion criteria, we found a moderate correlation between placebo response with publication year (MPC R²=0.33 and RR50 R²=0.36). When inclusion was based on history alone (N=9), there was a weak inverse correlation (R²=0.17) between RR50 and publication year. There was a weak inverse correlation (R²=0.12, N=20) between RR50 and duration of historical period. A weak inverse correlation between placebo outcomes (RR50 R²=0.13 and MPC R²=0.12) and minimum seizure frequency requirement during historical or baseline periods (N=46). _x000D_
Conclusions: There appears to be some subtle relationships between inclusion criteria and placebo response rates depending on the outcome metric (MPC and RR50). Further study may reveal optimal choices for RCT inclusion criteria to minimize placebo response rates._x000D_
Funding: Funding in part from NIH K23NS124656