Abstract
Byline: Ayham Alkhachroum, Univ of Miami, Miami, FL; Lili Zhou, Univ of Miami, Miami, FL; Negar Asdaghi, UNIVERSITY OF MIAMI MILLER SCHOOL, Miami, FL; Hannah Gardner, Univ of Miami, Miami, FL; Hao Ying, Univ of Miami, Miami, FL; Carolina M Gutierrez, UNIVERISTY OF MIAMI MILLER SCH, Miami, FL; Daniel Samano, Univ of Miami, Miami, FL; Danielle Bass, Univ of Miami, Miami, FL; Dianne Foster, American Heart Association, Atlanta, GA; Nicole B Sur, Univ of Miami Neurology, Miami, FL; Nina Massad, Univ of Miami, Miami, FL; Mohan Kottapally, Univ of Miami, Miami, FL; Amedeo Merenda, miami, FL; Robert Starke, Miami, FL; Kristine O'Phelan, Univ of Miami, Miami, FL; Jose G Romano, U Miami, Pinecrest, FL; Jan Claassen, Columbia Univ Med Cntr, New York, NY; Ralph L Sacco, UNIVERSITY OF MIAMI, Miami, FL; Tatjana Rundek, MILLER SCHL OF MEDICINE UNIV M, Miami, FL Introduction: The decision to withhold or withdraw life-sustaining therapy (WLST) is common after acute stroke. Factors that may influence the decision are not well determined. We aimed to investigate factors associated with WLST in hospitalized acute stroke patients. Methods: Patients with acute ischemic stroke (AIS), intracerebral hemorrhage (ICH), and subarachnoid hemorrhage (SAH) were included across 152 Florida hospitals participating in the prospective Florida Stroke Registry from 2008-2021. Importance plots were performed to generate the predictive factors associated with WLST. AUC-ROC curves were generated for the performance of logistic regression (LR) and random forest (RF) models. We used 75/15/15 for training/testing/validation. Results: Among 309,393 AIS patients, 47,485 ICH patients, and 16,694 SAH patients; 9%, 28%, and 19% subsequently had WLST during hospitalization. Patients who had WLST were older (77 vs. 69 years), more women (57% vs. 49%), more White (76% vs. 67%), greater stroke severity at presentation NIHSS [greater than or equal] 5 (29% vs.19%), more likely to be treated in comprehensive stroke centers (52% vs. 44%), more likely to have Medicare insurance (53% vs. 44%), less likely to be uninsured (8% vs. 13%), more likely to undergo surgical treatments (1.2% vs 0.3%), and more likely to have impaired level of consciousness (38% vs. 12%). The most predictive factors associated with the decision to WLST in AIS were age, stroke severity, state region, insurance status, stroke center type, race, and level of consciousness (RF AUC of .93 and LR AUC of .85). The most predictive factors in ICH were age, impaired level of consciousness, state region, race, insurance status, stroke center type, and ambulation status at baseline (RF AUC of .76 and LR AUC of .71). Most predictive factors in SAH were age, impaired level of consciousness, state region, insurance status, race, and stroke center type (RF AUC of .82 and LR AUC of .72). Conclusion: Among acute hospitalized stroke patients; age, level of consciousness, state region, race, insurance status, ambulation status at baseline, and stroke center type could contribute to the decision to WLST.
