Cerebral Arteriovenous Malformations and Epilepsy, Part 2: Predictors of Seizure Outcomes Following Radiosurgery

OBJECTIVE: Seizure outcomes after arteriovenous malformation (AVM) management with radiosurgery are incompletely understood. In this case-control study, we aim to determine the incidences and define the predictors of seizure improvement and de novo seizures in patients with AVM with and without seizures at presentation, respectively. METHODS: We evaluated our institutional AVM radiosurgery database to determine the factors that were associated with favorable seizure outcome (seizure improvement or lack of de novo seizures). In patients with seizures at presentation, seizure improvement was defined as decreased seizure frequency or complete seizure remission. In patients without seizures at presentation, de novo seizures were defined as new-onset seizures after radiosurgery. Logistic regression analyses were performed to identify predictors of favorable seizure outcome. RESULTS: In 229 patients with seizures at presentation, the rates of seizure improvement and seizure remission were 57% and 20%, respectively. Prior AVM hemorrhage (P = 0.015), longer follow-up (P < 0.0001), and lack of hemorrhage after radiosurgery (P = 0.048) were independent predictors of seizure improvement in the multivariate analysis. In 778 patients without seizures at presentation, the overall rate of de novo seizures was 1.7%. Prior AVM hemorrhage (P = 0.001) and higher Spetzler-Martin grade (P = 0.018) were independent predictors of the absence of de novo seizures in the multivariate analysis. AVM obliteration was not significantly associated with seizure outcomes after radiosurgery. CONCLUSIONS: Radiosurgery provides reasonable rates of seizure improvement for patients with AVM who present with seizures. For patients with AVM without seizures at presentation, the risk of de novo seizures after radiosurgery is very low, obviating the need for prophylactic antiepileptic drug therapy. Further investigation of epilepsy in patients with AVM undergoing stereotactic radiosurgery should be considered with validated outcome measures and prospective study design.