Abstract
Abstract only Introduction: Seizures are a common initial manifestation of cerebral venous thrombosis (CVT). We sought to establish predictive factors for delayed seizures to stratify risk and guide clinical management. Methods: We analyzed data from the Direct Oral Anticoagulants Versus Warfarin in the Treatment of Cerebral Venous Thrombosis (ACTION-CVT) study of consecutive patients with CVT from 2015 - 2020. The primary outcome was delayed seizure(s), occurring after 7 days from CVT diagnosis. Patients with outcome data for late seizures and follow-up of at least 7 days from time of diagnosis were included. Univariate analysis was performed using Chi-square, rank-sum, and t-tests. Variables with P-value < 0.2 were included in the backward stepwise regression model with 0.2 as cutoff, based on which we constructed the Delayed Seizure CVT (Delayed Seizures after CVT) score. Results: We identified 599 patients who met inclusion criteria: mean age 44.9(± 16.6) years, 62.9% female. Delayed seizures occurred in 37 (6.1%) patients. Risk factors included intracranial hemorrhage (OR 2.0 95% CI 0.7-5.74 p=0.196), alcohol abuse (OR 3.88 95% CI 1.45-10.38 p=0.007), focal neurologic deficit (OR 2.05 95% CI 0.95-4.44 p=0.067), seizure on presentation (OR 2.88 95% CI 1.32-6.29 p=0.008), hormonal birth control (OR 2.38 95% CI 1.06-5.37 p=0.0036), endovascular therapy (OR 2.86 95% CI 1.16-7.07 p=0.023) and decompressive hemicraniectomy (OR 3.357 95% CI 1.02-10.99 p=0.046), respectively. Based on coefficients of these variables, a 7-item score was created with acceptable performance for seizure prediction (AUC 0.796, 95% CI 0.72 - 0.86) (Table 1). Conclusions: Delayed seizure CVT score is a clinical tool to identify risk of delayed seizure in CVT. Further research is needed to validate these findings in an external cohort.