An Admission-Only Prognostic Model for Early Outcome Prediction After Aneurysmal Subarachnoid Haemorrhage: Development and Temporal Validation
Discuss this preprint
Start a discussion What are Sciety discussions?Listed in
This article is not in any list yet, why not save it to one of your lists.Abstract
Background
Early prognostic estimation after aneurysmal subarachnoid haemorrhage (aSAH) is often required before complete aneurysm characterisation, neurological reassessment and treatment-related information are available. In this study, we developed and temporally validated an admission-only model for favourable 90-day functional outcomes.
Methods
We retrospectively analysed consecutive treatment-eligible patients with aSAH who underwent aneurysm-securing treatment between 2021 and 2025. Patients treated during 2021–2023 and 2024–2025 formed the development and temporal validation cohorts, respectively. Predictors were age, pre-morbid modified Rankin Scale (mRS), admission World Federation of Neurosurgical Societies grade, modified Fisher grade and intracerebral haemorrhage on initial computed tomography. Primary and sensitivity outcomes were 90-day mRS 0–3 and 0–2, respectively. Model performance was assessed using discrimination, Brier score, calibration and decision curve analysis, with contextual benchmarking against admission WFNS grade alone and the SAFIRE grading scale.
Results
Overall, 245 patients were included (development cohort, n=160; validation cohort, n=85). In temporal validation, AUCs were 0.868 for mRS 0–3 and 0.870 for mRS 0–2; corresponding Brier scores were 0.134 and 0.147. The admission-only model showed higher discrimination and lower Brier scores than admission WFNS grade alone. The SAFIRE-based contextual benchmark yielded higher performance but required information beyond initial presentation. Exploratory external assessment showed AUCs of 0.790 and 0.775 for 6-month mRS 0–3 and 0–2, respectively.
Conclusions
This admission-only model may complement established prognostic scores by supporting early prognostic estimation before all components of the SAFIRE grading scale are available.
KEY MESSAGES
WHAT IS ALREADY KNOWN ON THIS TOPIC
Aneurysmal subarachnoid haemorrhage is a severe stroke subtype associated with substantial mortality and long-term disability. Existing prognostic scores, including SAFIRE, provide useful risk stratification, but some components may not be uniformly available at the initial presentation before complete aneurysm characterisation and neurological reassessment.
WHAT THIS STUDY ADDS
In this study, we developed and temporally validated a simple admission-only prognostic model using five readily available variables: age, premorbid modified Rankin Scale score, admission World Federation of Neurosurgical Societies grade, modified Fisher grade, and intracerebral haemorrhage on initial CT. The model provides individualised probabilities of favourable 90-day functional outcomes before treatment-related information is available, and was evaluated using discrimination, calibration, decision curve analysis, and exploratory external assessment.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
By focusing on the earliest clinical time point, this model may help bridge the gap between initial presentation and complete prognostic assessment using established scores.
