The prognostic impact of epilepsy in glioblastoma: Insights from a systematic review and meta-analysis.

Abstract

e14038 Background: Epilepsy is a common comorbidity in patients with glioblastoma (GBM) and significantly impacts quality of life and clinical management. Whether epilepsy itself or the choice of antiepileptic drugs (AEDs) influences overall survival remains controversial, with conflicting results reported across observational and interventional studies. Methods: Following PRISMA 2020 guidelines, a systematic review and meta-analysis were performed across PubMed, Cochrane Library, Google Scholar, Mendeley, and Open Evidence (January 2015 - December 2025). Eligible studies included adults (≥18 years) with histologically confirmed GBM reporting seizure incidence, clinical features, or treatment outcomes. Data extraction and bias assessment were independently conducted by four reviewers. Results: Thirty-four studies were included, comprising 28 primary studies and 6 prior systematic reviews. Across high-quality pooled and registry-based analyses, no consistent survival advantage was associated with the use of specific AEDs, including valproic acid or levetiracetam, after adjustment for molecular and clinical confounders. Levetiracetam demonstrated equivalent or superior seizure control with a more favorable safety profile compared with valproic acid. IDH mutation status and MGMT promoter methylation emerged as the strongest independent prognostic factors for both survival and seizure-related outcomes. Conclusions: Epilepsy in GBM is associated with favorable molecular subtypes and improved survival, likely reflecting less aggressive tumor biology. Levetiracetam remains the preferred agent for tumor-related epilepsy management in glioblastoma. Quality assessment of observational studies (Newcastle-Ottawa Scale). Study Selection (Max 4) Comparability (Max 2) Outcome (Max 3) Total Score (Max 9) Overall Quality Knudsen-Baas et al. (2016) 4 2 3 9 Very Low Risk Chen et al. (2017) 4 2 2 8 Low Risk Watanabe et al. (2017) 4 1 3 8 Low Risk Toledo et al. (2017) 3 2 3 8 Low Risk Salvati et al. (2018) 3 2 3 8 Low Risk Pallud et al. (2024) 3 2 3 8 Low Risk Bianconi et al. (2024) 3 2 3 8 Low Risk Cardona et al. (2017) 3 2 3 8 Low Risk Yu et al. (2018) 3 2 3 8 Low Risk Happold et al. (2016) 3 2 3 8 Low Risk Salvati et al. (2020) 3 2 2 7 Low Risk Feyissa et al. (2019) 3 2 2 7 Low Risk Li et al. (2021) 3 2 2 7 Low Risk Bushara et al. (2021) 3 2 2 7 Low Risk Ollila & Roivainen (2025) 3 1 3 7 Low Risk Al-Dorzi et al. (2023) 3 2 2 7 Low Risk Ryu et al. (2019) 3 2 2 7 Low Risk Tinchon et al. (2015) 3 2 2 7 Low Risk Yang et al. (2023) 4 1 2 7 Low Risk Hattori et al. (2023) 3 2 2 7 Low Risk van der Meer et al. (2023) 3 2 2 7 Low Risk Królicki et al. (2018) 3 0 3 6 Moderate Risk Sughrue et al. (2015) 3 1 2 6 Moderate Risk Casas Parera et al. (2019) 3 0 3 6 Moderate Risk Maschio et al. (2017) 2 0 2 4 Moderate Risk Rades et al. (SURF-ROGG) 1 0 1 2 High Risk