Abstract:
Stroke remains a leading global cause of mortality and disability, necessitating effective tools for early risk
stratification. This systematic review evaluates the prognostic performance of Early Warning Scores
(EWS)—MEWS, NEWS, and NEWS2—in predicting in-hospital mortality among stroke patients. Five studies
published between 2021 and 2025 were included, encompassing diverse stroke subtypes and care settings.
Across studies, higher EWS values were consistently associated with increased mortality risk, with odds
ratios ranging from 2.44 to 3.21 and AUROC values between 0.69 and 0.79. NEWS2 generally outperformed
MEWS and NEWS. Importantly, models incorporating neurological assessments, particularly the Glasgow
Coma Scale (GCS), significantly improved predictive accuracy (AUROC 0.87). While EWS are valuable for
early detection of deterioration, their performance in stroke is enhanced by integrating neurological
parameters. These findings support the development of combined scoring systems and digital integration
to optimize early intervention strategies in acute stroke care across varied healthcare settings
