Predicting the occurrence of early seizures after cerebral venous thrombosis using a comprehensive nomogram.

ABSTRACT

BACKGROUND: Seizure is a common clinical manifestation of cerebral venous thrombosis (CVT). The mortality rate of patients with CVT with seizure is three times higher than that of patients without seizure. The aim of this study was to develop a nomogram to predict the individual probability of acute seizure events in patients with CVT. METHOD: This was a single-center, retrospective cohort study. We analyzed and compared demographic variables, epidemiological risk factors, clinical presentation, laboratory results and imaging data in a cohort of 142 patients who were diagnosed with CVT in our hospital from January 2013 to December 2018. A nomogram was constructed to predict the risk of early seizure (ES) in these patients according to the multivariable logistic regression analysis results. The concordance index, GiViTi calibration belt and decision curve analysis (DCA) were used to assess nomogram performance. RESULTS: Forty-three (30.28%) patients experienced seizure within 2 weeks after a CVT diagnosis. Multivariate analysis identified focal neurologic deficit, Glasgow Coma Scale (GCS) scores ≤ 8 on admission, hemorrhagic lesions, superior sagittal sinus thrombosis (SSST) and frontal lobe lesions as independent predictive factors for ES occurrence after CVT. A nomogram was generated based on these predictive factors with the concordance index reaching 0.82, indicating that the clinical tool was well calibrated. DCA showed that the model was useful with a threshold probability in the range of 0-77%. CONCLUSIONS: We developed the first nomogram that could predict the risk of ES in CVT patients. This effective and convenient tool has shown promising clinical benefit and will assist clinicians in making treatment decisions.