Epilepsy in Patients With Primary CNS Lymphoma: Prevalence, Risk Factors, and Prognostic Significance

• Published in: Neurology Vol. 103; no. 5; p. e209748
• Main Authors: Aboubakr, Oumaima, Houillier, Caroline, Alentorn, Agusti, Choquet, Sylvain, Dupont, Sophie, Mokhtari, Karima, Leclercq, Delphine, Nichelli, Lucia, Kas, Aurelie, Rozenblum, Laura, Le Garff-Tavernier, Magali, Hoang-Xuan, Khê, Carpentier, Alexandre, Mathon, Bertrand
• Format: Journal Article
• Language: English
• Published: United States 10.09.2024
• Subjects: Adult; Aged; Aged, 80 and over; Central Nervous System Neoplasms - complications; Central Nervous System Neoplasms - epidemiology; Epilepsy - epidemiology; Female; Humans; Lymphoma - complications; Lymphoma - epidemiology; Male; Middle Aged; Prevalence; Prognosis; Retrospective Studies; Risk Factors
• ISSN: 1526-632X
• DOI: 10.1212/WNL.0000000000209748

Epilepsy is a common comorbidity of brain tumors; however, little is known about the prevalence, onset time, semiology, and risk factors of seizures in primary CNS lymphoma (PCNSL). Our objectives were to determine the prevalence of epilepsy in PCNSL, to identify factors associated with epilepsy, and to investigate the prognostic significance of seizures in PCNSL. We performed an observational, retrospective single-center study at a tertiary neuro-oncology center (2011-2023) including immunocompetent patients with PCNSL and no history of seizures. We collected clinical, imaging, and treatment data; seizure status over the course of PCNSL; and oncological and seizure outcome. The primary outcome was to determine the prevalence of epilepsy. Furthermore, we aimed to identify clinical, radiologic, and treatment-related factors associated with epilepsy. Univariate analyses were conducted using the χ test for categorical variables and unpaired test for continuous variables. Predictors identified in the unadjusted analysis were included in backward stepwise logistic regression models. We included 330 patients, 157 (47.6%) were male, median age at diagnosis was 68 years, and the median Karnofsky Performance Status score was 60. Eighty-three (25.2%) patients had at least 1 seizure from initial diagnosis to the last follow-up, 40 (12.1%) as the onset symptom, 16 (4.8%) during first line of treatment, 27 (8.2%) at tumor progression and 6 (1.8%) while in remission. Focal aware seizures were the most frequent seizure type, occurring in 43 (51.8%) patients. Seizure freedom under antiseizure medication was observed in 97.6% patients. Cortical contact (odds ratio [OR] 8.6, 95% CI 4.2-15.5, < 0.001) and a higher proliferation index (OR 5.7, 95% CI 1.3-26.2, = 0.02) were identified as independent risk factors of epilepsy. Patients with PCNSL and epilepsy had a significantly shorter progression-free survival (median progression-free survival 9.6 vs 14.1 months, adjusted hazard ratio 1.4, 95% CI 1.0-1.9, = 0.03), but not a significantly shorter overall survival (17 vs 44.1 months, log-rank test, = 0.09). Epilepsy affects a quarter of patients with PCNSL, with half experiencing it at the time of initial presentation and potentially serving as a marker of disease progression. Further research is necessary to assess the broader applicability of these findings because they are subject to the constraints of a retrospective design and tertiary center setting.