Development and Temporal Validation of a Simple Admission-Based Model Predicting 90-Day Outcomes After Aneurysmal Subarachnoid Hemorrhage

BACKGROUNDEarly prognostic stratification of aneurysmal subarachnoid hemorrhage (aSAH) is clinically important. We developed and temporally validated an admission-based prediction model for 90-day outcomes in a cohort of patients with aSAH. METHODSConsecutive treatment-eligible patients with aSAH, managed at a single center between January 2021 and December 2025, were retrospectively analyzed. Patients treated from 2021-2023 and 2024-2025 comprised the development and temporal validation cohorts, respectively. Prediction models were developed using admission variables, including age, pre-morbid modified Rankin Scale (mRS) score, World Federation of Neurosurgical Societies grade, modified Fisher grade, and intracerebral hemorrhage on initial computed tomography. The primary outcome was a 90-day mRS score of 0-3. A sensitivity model was constructed for 90-day mRS scores of 0-2. RESULTSA total of 245 patients were included (development cohort: 160; validation cohort: 85); 107 patients were aged [≥] 70 years. For mRS 0-3, the model demonstrated good discrimination in both the development and validation cohorts (area under the curve [AUC]: 0.917 and 0.868), while mRS 0-2 had corresponding AUCs of 0.920 and 0.840, respectively. Among patients aged [≥] 70 years, the validation AUCs were 0.842 and 0.768 for mRS scores of 0-3 and 0-2, respectively. Calibration was acceptable overall but less stable in older patients. CONCLUSIONSIn an aSAH cohort with a substantial proportion of older patients, an admission-based five-variable model provided 90-day outcome prediction with good discrimination on temporal validation. This tool may facilitate early risk stratification at admission, pending multicenter external validation. CLINICAL PERSPECTIVEO_ST_ABSWhat is New?C_ST_ABSO_LIWe developed and temporally validated a simple admission-based prediction model for 90-day functional outcomes after aneurysmal subarachnoid hemorrhage in a contemporary, consecutive, single-center cohort comprising a substantial proportion of older patients. C_LIO_LIBased on five readily available admission variables, the model showed good discrimination in both the overall population and the subgroup aged [≥] 70 years, with additional support for clinical utility from decision curve analysis and an exploratory external assessment in an independent published cohort. C_LI What are the Clinical Implications?O_LIAdmission-based prediction may support early prognostic estimation, family counseling, and initial risk stratification at admission when treatment-related information is unavailable. C_LIO_LIBecause the cohort comprised treatment-eligible patients who received active aneurysm treatment, the model was intended for prognostic estimation rather than for guiding treatment allocation or withdrawal decisions. C_LI