Time-dependent predictors of seizure outcome in grade 1-2 glioma-related epilepsy

Given the reciprocal interaction between tumor biology and seizure activity, seizure outcomes in low-grade glioma (LGG) may be dynamic and influenced by both tumor- and treatment-related factors. We aimed to identify factors associated with seizure occurrence across distinct clinical stages, including diagnosis, surgery, long-term follow-up, and antiseizure medication (ASM) withdrawal.

We retrospectively analyzed patients with World Health Organization grade 1-2 glioma who underwent surgery between January 2001 and February 2025 and experienced seizures during their disease course. Clinical, radiological, molecular, and treatment-related variables, along with longitudinal seizure data, were collected. Seizure outcomes were evaluated at 6 and 12 months and at final follow-up, including time to postoperative seizure recurrence and recurrence after ASM withdrawal.

Among 100 patients (mean postoperative follow-up = 89.9 months), 70.0% achieved seizure freedom during the final year of follow-up. At 6 months, gross total resection (odds ratio [OR] = 3.45, 95% confidence interval [CI] = 1.19-10.01, p = .02) and preoperative tumor volume (OR = .99, 95% CI = .97-1.00, p = .048) were independently associated with seizure outcomes. At 12 months, preoperative tumor volume (OR = .99, 95% CI = .98-1.00, p = .04) was significant. Conversely, seizure at presentation was the sole independent determinant of favorable long-term seizure outcome (OR = 6.37, 95% CI = 2.07-19.61, p < .01) and the only predictor of reduced postoperative seizure recurrence in survival analysis (hazard ratio = .26, 95% CI = .11-.62, p < .01). ASM withdrawal was successful in 67.7%. Although tumor progression showed a trend toward an association with recurrence in univariate analyses, no independent predictors were identified in Cox regression for ASM withdrawal outcomes.

This time-resolved analysis suggests that early seizure control is primarily influenced by surgical and tumor burden-related factors, whereas long-term seizure outcomes and recurrence appear to be predominantly determined by the clinical presentation at diagnosis rather than treatment-related variables. These findings suggest that postoperative seizure prognosis in LGG reflects pre-existing epileptogenic vulnerability and may not be adequately captured by single time point assessments.