Clinical factors associated with seizure freedom in patients with focal-onset seizures (FOS) receiving perampanel 4 mg/day in FREEDOM study 342

Abstract

Purpose: Previous analyses have identified factors that affect treatment responses, e.g. high baseline seizure frequency has been associated with poor prognosis (French JA. Epilepsy Curr 2002;2:69–71). This post hoc analysis identified predictive clinical factors for patients (aged 12–74 years) with FOS, with/without focal to bilateral tonic-clonic seizures (FBTCS), achieving 26 weeks’ seizure freedom while receiving perampanel 4 mg/day during FREEDOM Study 342 (NCT03201900).

Method: This analysis was based on the 4-mg/day Treatment Phase (6-week Titration; 26-week Maintenance). Logistic regression was used to determine odds ratios (ORs) and 95% confidence intervals (CIs). Goodness of fit was measured by area under the receiver operating characteristic (ROC) curve. Baseline seizure frequency, seizure history, age at diagnosis, time since diagnosis, etiology and perampanel plasma concentration were assessed individually as continuous and/or categorical variables in univariate analyses. All variables were included in multivariate analyses with forwards/backwards/no selection. Patients who withdrew during Titration were excluded.

Results: Overall, 46/73 (63.0%) patients with ≥1 post-dose efficacy measurement remained seizure free for 26 weeks on perampanel 4 mg/day. In univariate analyses, lower baseline seizure frequency was a predictor of seizure freedom as both a continuous (P = 0.0136) or categorical (≤2 vs >2 seizures/12 weeks; P = 0.0284) variable. History of focal aware (P = 0.7137), focal impaired awareness (P = 0.3711) or FBTCS (P = 0.7005); age at diagnosis (P = 0.9241); time since diagnosis (P = 0.1862); idiopathic (P = 0.7118) or structural (P = 0.9033) etiology; and perampanel plasma concentration (P = 0.3940) were not predictors of seizure freedom. In multivariate analyses (no selection), only lower baseline seizure frequency was a predictor of seizure freedom (OR [95% CI], 0.817 [0.696, 0.960]; P = 0.0140; area under ROC curve, 0.7483). The same model was selected using forwards/backwards selection.

Conclusion: The best predictor for seizure freedom was baseline seizure frequency; the lower the baseline seizure frequency, the higher the probability of remaining seizure free.

Funding: Eisai Co., Ltd