Abstract TP110: Incident ICU Transfer Characteristics Among Ischemic Stroke Patients: A 4-year Retrospective Analysis

Abstract

Background: Neurologic patients transferred to an intensive care unit (ICU) have increased length of stays, mortality, and healthcare cost. There is limited data regarding incident transfers in stroke-specific cohorts. We sought to identify clinical features, transfer timing, and outcomes of patients admitted with ischemic stroke who required ICU escalation. Methods: We identified all patients age >18 admitted to our institution with a principle diagnosis of ischemic stroke between 2016 and 2019. We included patients requiring transfer to an ICU at any point in their hospitalization. We identified indications for transfer (IFT), time to ICU transfer (TTI), and outcomes at discharge and at 90-days post-discharge. We subdivided the IFT into neurologic and non-neurologic etiologies, and used chi-square testing and Kaplan-Meier curves with log-rank modeling to identify differences between patients with neurologic and non-neurologic transfer indications. Results: 3176 patients were admitted with a principle diagnosis of ischemic stroke across the four-year study period. 118 stroke patients (3.7%) underwent incident ICU transfer, 49 (41.5%) due to neurological causes and 69 (58.5%) due to non-neurological causes. The most common neurologic IFTs were cerebral edema (51.0%) and intracerebral hemorrhage (14.3%). The most common non-neurologic IFTs were respiratory decline (47.8%) and shock (21.7%). There were no differences in stroke characteristics (last known normal prior to arrival, baseline NIHSS, thrombolytic rates) between patients with neurologic vs. non-neurologic IFTs. Patients with neurologic IFTs had shorter mean TTI (2.45 vs. 4.80 days; p = 0.035) and were less likely to be discharged home compared to patients with non-neurologic IFTs (21.7% vs. 6.1%; p = 0.021). We corroborated differences in IFTs by adjusted log-rank comparisons on survival analysis. There were no significant difference in length of ICU stay (4.66 vs. 4.83 days) or total length of stay (12.37 vs 16.75 days). Conclusions: Most ischemic stroke patients’ incident ICU transfers occur in the first 48 hours, highlighting the need for early vigilance. Neurologic and non-neurologic etiologies for IFT may contribute to the patients’ disposition outcome.