To summarize evidence on the effectiveness of the SeLECT score in predicting the risk of post-stroke epilepsy (PSE).

A systematic search was carried out in PubMed, Embase, The Cochrane Library, and Web of Science (Clarivate Analytics) using a predefined search criterion for relevant literature from inception to October 2023. Two authors independently performed a database search, data collection, and study quality evaluation. The primary outcome of the study was late seizures after stroke.

This review included data from 6 patient cohorts with more than 3400 stroke patients from 6 countries (Switzerland, Austria, Germany, Italy, China and Turkey).

SeLECT was a significant predictor of PSE within one year (HR = 1.8, 95% CI 1.6–2.1, p <0·0001). The AUC of the SeLECT score was 0.756 (SE = 0.033, 95% CI = 0.692–0.819). Interestingly, when combined with the biomarker IL-1β, a critical inflammatory cytokine contributing to neuroinflammation after ischemic stroke, the AUC of the combination went up to 0.933 (SE = 0.027, 95% CI = 0.880–0.985) with a sensitivity of 88% and specificity of 82%.

In patients receiving IV thrombolytic therapy, adding two additional variables to the score, namely diabetes and leukoaraiosis, improved its ability to predict PSE (AUC = 0.955 vs 0.893) with a sensitivity of 89% and specificity of 93%.

SeLECT score is an effective and easy-to-use clinical tool that accurately predicts the risk of PSE. However, it is crucial to validate its effectiveness in the United States population. Given the lack of reliable interventions to identify high-risk patients for PSE, we believe that this study presents an opportunity for larger multicenter studies, particularly in the United States, to develop precise interventions that can prevent this common yet terrifying complication of stroke.