Abstract WP012: Specific Evidence-Based Diagnostic Tests are Associated with Greater Likelihood of Identifying Non-Cryptogenic Ischemic Stroke Etiology at Discharge, Yet Cryptogenic Strokes Remain: New Results from the American Heart Association and

BACKGROUND: About one-fourth of ischemic strokes are recurrent. Finding a specific etiology facilitates targeted secondary prevention. The American Heart Association (AHA) Get-With-The-Guidelines-Stroke® Registry (GWTG-S) suggests more than 50% of patients are discharged with a cryptogenic or undocumented etiology. We sought to determine baseline characteristics associated with cryptogenic stroke and whether more thorough hospital investigation reduces cryptogenic classification. METHODS: In 2022, the AHA, along with HCA Healthcare and HCA Healthcare Foundation, designed and launched Getting to the Heart of Stroke TM (GTTHOS) in 10 HCA Healthcare advanced stroke centers to improve evidence-based stroke evaluation, promote neurology/cardiology collaboration, and better identify barriers to care. GTTHOS data from 2022-2024 were harvested from GWTG-S, noting baseline characteristics (demographics, insurance status, medical history, NIHSS) and utilization of three diagnostic tests: cervical carotid imaging, intracranial vascular imaging, and short-term cardiac rhythm monitoring. Multivariate analysis sought to associate baseline features and test utilization with cryptogenic vs. non-cryptogenic etiology. RESULTS: We identified 15,216 patients with cryptogenic (5,468, 35.94%) or non-cryptogenic (9,748, 64.06%) ischemic stroke. Cryptogenic patients had lower median age (69 vs. 72 years, p < 0.0001), more frequent vascular risk factors (hypertension, diabetes, dyslipidemia), and lower prevalence of heart failure or coronary disease (Table 1). Baseline differences were small and mostly nonsignificant when adjusted for covariates. Performing diagnostic tests was associated with higher likelihood of identifying a non-cryptogenic etiology (Table 2): intracranial vascular imaging (aOR = 1.24, p = 0.0146), carotid imaging (aOR = 1.67, p < 0.0001), and short-term cardiac rhythm monitoring (aOR = 1.23, p = 0.0004). Test utilization was highest in cryptogenic stroke. CONCLUSIONS: Performance of diagnostic tests (intracranial vascular imaging, carotid imaging, and short-term cardiac monitoring) was strongly associated with greater likelihood of identifying a non-cryptogenic stroke etiology. Cryptogenic patients were thoroughly, though not exhaustively, assessed with these tests, making it unlikely that cryptogenic strokes were simply under-tested. Opportunities exist for improved adherence to AHA acute and post-acute testing guidelines.