PC1014: Variability in Diagnostic Imaging for Children Presenting to the Emergency Department with Acute Stroke

Background: Pediatric stroke carries significant morbidity and mortality, compounded by frequent delays in diagnosis and treatment. Current guidelines support the acquisition of MRI as first-line imaging in children with suspected acute thrombotic stroke as well as early acquisition of head angiography to improve diagnostic accuracy and mitigate delays in diagnosis and treatment. It is unknown which patient characteristics may affect acquisition of definitive imaging in the emergency department (ED) among children with stroke.

Methods: We performed a retrospective cohort study of children using administrative data from 44 geographically diverse US children’s hospitals. We identified patients 29 days through 17 years of age with an ED encounter and an “International Classification of Disease, 10th revision” diagnosis code for acute thrombotic stroke (I63, H34.1) or transient ischemic attack (TIA) (G45) between October 1, 2015 and December 31, 2021. We excluded patients with a principal diagnosis code of trauma (T) or external injury (S), patients who did not receive head imaging on day of presentation, and patients discharged from the ED. Our primary outcomes were acquisition of MRI (versus CT only) and head angiography (versus no angiography) on day of ED presentation. We performed a generalized linear mixed model with binomial distribution to determine the association of illness severity (intensive care unit and mechanical ventilation), demographics (age, sex, insurance, and race/ethnicity), operational characteristics (urbanicity, outside hospital transfers, and presenting pre-2019), and pre-existing conditions with our outcomes of interest, with hospital as a random effect.

Results: We included 2,045 patients (Table 1). In multivariable analysis, factors associated with lower odds of MRI acquisition on day of presentation included young age ( < 12 months, 12-35 months), mechanical ventilation, and previous history of vascular malformation (VM) and central nervous system (CNS) cancer (Table 2). Factors associated with lower odds of angiography included Black race and other non-Hispanic non-white races, young age ( < 12 months, 12-35 months, 3-5 years), mechanical ventilation, and previous history of CNS cancer, metabolic disorders, thrombotic stroke, and seizures.

Conclusion: Among children presenting to the ED with acute thrombotic stroke and TIA, younger and non-white patients had lower odds of prompt definitive imaging. These findings may potentially be driven by system limitations for rapid MRI and delays in recognition, highlighting a need for clinical care pathways and rapid MRI protocols to be equitable and inclusive of higher-risk and young children. Furthermore, several pre-existing conditions appeared to modify the odds of MRI and angiography, which may imply nuances in clinical practice reflective of underlying stroke pathophysiology. We suggest that imaging tailored to comorbidities should be reflective of current evidence and standardized across clinical care pathways.