Background: Symptomatic epilepsy is a common symptom of glioblastoma, which may occur in different stages of disease. There are discrepant reports on association between early seizures and glioblastoma survival, even less is known about the background of these seizures. We aimed at analyzing the risk factors and clinical impact of perioperative seizures in glioblastoma.

Methods: All consecutive cases with de-novo glioblastoma treated at our institution between 01/2006 and 12/2018 were eligible for this study. Perioperative seizures were stratified into seizures at onset (SAO) and early postoperative seizures (EPS, ≤21days after surgery). Associations between patients characteristics and overall survival (OS) with SAO and EPS were addressed.

Results: In the final cohort (n = 867), SAO and EPS occurred in 236 (27.2%) and 67 (7.7%) patients, respectively. SAO were independently predicted by younger age (P = .009), higher KPS score (P = .002), tumor location (parietal lobe, P = .001), GFAP expression (≥35%, P = .045), and serum chloride at admission (>102 mmol/L, P = .004). In turn, EPS were independently associated with tumor location (frontal or temporal lobe, P = .013) and pathologic laboratory values at admission (hemoglobin < 12 g/dL, [P = .044], CRP > 1.0 mg/dL [P = 0.036], and GGT > 55 U/L [P = 0.025]). Finally, SAO were associated with gross-total resection (P = .006) and longer OS (P = .030), whereas EPS were related to incomplete resection (P = .005) and poorer OS (P = .009).

Conclusions: In glioblastoma patients, SAO and EPS seem to have quite different triggers and contrary impact on treatment success and OS. The clinical characteristics of SAO and EPS patients might contribute to the observed survival differences.