The SeLECT score, comprised of clinical variables including stroke severity, large-artery atherosclerotic etiology, early seizures, cortical involvement, and territory of middle cerebral artery involvement, is a feasible and effective tool for predicting the risk for late seizures occurring >7 days after an ischemic stroke, according to study findings published in Lancet Neurology.
Using a prospective registry of post-stroke seizures experienced by patients with ischemic stroke (n=1200) from a regional acute neurology center in Switzerland, investigators developed a model for predicting late seizure risk (>7 days after stroke) based on 5 key clinical variables: stroke severity, large-artery atherosclerotic etiology, early seizures, cortical involvement, and territory of middle cerebral artery involvement (SeLECT score). A validation study of this score was applied to participants with ischemic stroke from Austria (n=459), Germany (n=311), and Italy (n=399). On average, complete data were obtained for 99.2% of the risk predictors.
Approximately 1 year after stroke, the risk of late seizures was 4% (95% CI, 4-5) and at 5 years after stroke, the risk was 8% (95% CI, 6-9). A pooled data analysis of the 3 validation cohorts demonstrated that the SeLECT score was a significant predictor of late seizures (hazard ratio [HR] 1.8 per point; 95% CI, 1.6–2.1; P <.0001). The highest SeLECT score of 9 points predicted up to 63% (95% CI, 42-77) of late seizures within 1 year after stroke, whereas the lowest score of 0 points predicted 0.7% late seizure risk. At 5 years following stroke, the highest SeLECT score predicted approximately 83% (95% CI, 62-93) of late seizures. A concordance statistic of 0.77 (95% CI, 0.71-0.82) represented the model performance in the overall validation cohort. Based on calibration plots for the score, the investigators discovered a high agreement between predicted and the observed seizure risk.
Although this study supports the rationale for the SeLECT score for predicting post-stroke seizures, the findings may only be applied solely to ischemic strokes. Also, this study may have possessed some selection bias as the patients with the most severe strokes often died or were lost to follow-up.
The SeLECT score appears to be an effective risk prediction tool that may improve personalized medicine and “identify individuals who are most likely to benefit from antiepileptogenic interventions,” the investigators concluded.
Galovic M, Döhler N, Erdélyi-Canavese B, et al. Prediction of late seizures after ischaemic stroke with a novel prognostic model(the SeLECT score): a multivariable prediction model development and validation study. Lancet Neurol. 2018;17(2):143-152.