gms | German Medical Science

Objective: Glioblastoma (GBM) is the most common and aggressive primary brain tumor. In a high percentage of patients seizures occur as an initial symptom. Controversy exists whether seizures are primarily triggered by the site of the tumor or associated with imaging characteristics on initial MRI. Moreover, the influence of seizures on overall survival (OS) has rarely been evaluated in outcome analysis. The goal of this study was to identify potential triggers and confounders of seizures as a pretreatment symptom of GBM and to evaluate their impact on OS.

Methods: Patients harboring a primary GBM treated at two neurosurgical departments (from 2009 - 2016) with an initial high-quality MRI were included into this retrospective study. Clinical and follow-up data was collected from patient records and volumetric measurement of tumor compartments was performed in a 3D semi-automated fashion (tumor, necrosis, peritumoral edema). Focal and generalized seizures were subsumed. For survival analysis, a subpopulation was composed including patients treated with a surgical resection (maximum residual tumor volume of ≤ 2 cm3) followed by standardized concomitant radio- and chemotherapy. Uni- and multivariate models were calculated for confounders of onset of seizures and OS.

Results: A total of 224 patients were included in our bi-center study (OS subpopulation n= 99). Seizures occurred in 17,9% of all patients as an initial symptom. Younger patients (≤ 60y) were predominantly affected by seizures (p = 0,022) independent of the affected lobe. In volumetric analysis, all volumes were negatively correlated with the onset of seizures (p = 0,001), being not influenced by patient age. Statin medication was associated with significantly less seizures in uni- and multivariate analysis (p = 0,025, p = 0,004, respectively). OS was increased within the subpopulation of patients with initial seizures, without reaching statistical significance (p = 0,356). Statins had no effect on OS (p = 0,506).

Conclusion: Our study evaluated confounders for the appearance of seizures in primary GBM and analyzed the association with OS. Younger patients were most affected by seizures, while statins were protective towards their appearance. Neither seizures, nor statins had a significant effect on OS. This might be explained by young patient age outweighing the effect of seizures on OS. The volumes of all measured tumor compartments correlated with the occurrence of pretreatment seizures inversely.