Diagnostic accuracy of 3-item stroke scale for detection of cerebral large vessel occlusion: A systematic review and meta-analysis

Abstract

BackgroundPrompt identification of large vessel occlusion (LVO) in acute ischemic stroke (AIS) is crucial for expedited endovascular therapy (EVT) and improved patient outcomes. Prehospital stroke scales, such as the 3-Item Stroke Scale (3I-SS), could be beneficial in detecting LVO in suspected patients. This meta-analysis evaluates the diagnostic accuracy of 3I-SS for LVO detection in AIS. MethodsA systematic search was conducted in Medline, Embase, Scopus, and Web of Science databases until February 2024 with no time and language restrictions. Prehospital and in-hospital studies reporting diagnostic accuracy were included. Review articles, studies without reported 3I-SS cut-offs, and studies lacking the required data were excluded. Pooled effect sizes, including area under the curve (AUC), sensitivity, specificity, diagnostic odds ratio (DOR), positive and negative likelihood ratios (PLR and NLR) with 95% confidence intervals (CI) were calculated. ResultsTwenty-two studies were included in the present meta-analysis. A 3I-SS score of 2 or higher demonstrated sensitivity of 76% (95% CI: 52%–90%) and specificity of 74% (95% CI: 57%–86%) as the optimal cut-off, with an AUC of 0.81 (95% CI: 0.78–0.84). DOR, PLR, and NLR, were 9 (95% CI: 5–15), 2.9 (95% CI: 2.0–4.3) and 0.32 (95% CI: 0.17–0.61), respectively. Sensitivity analysis confirmed the analyses' robustness in suspected to stroke patients, anterior circulation LVO, assessment by paramedics, and pre-hospital settings. Meta-regression analyses pinpointed LVO definition (anterior circulation, posterior circulation) and patient setting (suspected stroke, confirmed stroke) as potential sources of heterogeneity. Conclusion3I-SS demonstrates good diagnostic accuracy in identifying LVO stroke and may be valuable in the prompt identification of patients for direct transfer to comprehensive stroke centers.