External multicenter validation of the eSAH score for predicting outcomes after subarachnoid hemorrhage

Reliable assessment of prognosis after aneurysmal subarachnoid hemorrhage (SAH) is essential to inform clinical decision-making and prevent premature assumptions of poor outcome. The enhanced SAH (eSAH) score combines three routinely available variables-age, Glasgow Coma Scale (GCS), and subarachnoid hemorrhage blood volume (SAHV)-to provide an objective framework for early outcome stratification. SAHV is estimated using the ABC/2-derived ellipsoid method applied across five major cisternal compartments on non-contrast CT. While the eSAH score has shown promise in initial derivation studies, its external performance and calibration have not yet been validated across diverse healthcare settings. We conducted a retrospective cohort study using consecutive data from two high-volume tertiary centers: one in a high-income country (the United States) and one in a middle-income country (Brazil). Non-aneurysmal SAH cases were excluded. The eSAH score was validated to predict discharge functional outcomes (modified Rankin Scale, mRS) and mortality, using discrimination (area under the ROC curve, AUC) and calibration metrics and was compared to the World Federation of Neurosurgical Societies (WFNS) score. A total of 472 patients were included. The eSAH score demonstrated moderate discriminative performance, with AUCs of 0.78 for unfavorable functional outcome and 0.799 for in-hospital mortality. Each one-point increase in the score was associated with a 2.14-fold increase in the odds of poor outcome (95% CI, 1.84-2.50; p < 0.0001) and a 2.14-fold increase in the odds of mortality (95% CI, 1.78-2.57; p < 0.001). The WFNS score yielded slightly lower AUCs (0.76 and 0.75, respectively). The eSAH score provides a simple, generalizable framework for early prognostication after aneurysmal SAH, performing consistently across two markedly different healthcare systems. Although its discriminative power remains moderate, its objective structure and reliance on measurable imaging features make it ideally suited for integration with artificial intelligence to enable automated, precise, and scalable SAHV quantification. Such integration may substantially enhance prognostic accuracy and support more equitable, data-driven neurocritical care worldwide.