45 External Validation of a Subdural Hematoma Risk Decision Instrument in Emergency Department Patients

Abstract

Anticoagulation, hematoma size, and midline shift are significant prognostic factors for poor outcomes in the acute phase of subdural hematomas (SDH). They were used by Pruitt et al in 2017 to develop a decision instrument identifying low risk patients with traumatic SDH. We wished to externally validate this decision instrument in our emergency department (ED), an academic, Level 1 trauma center. We conducted a retrospective chart review study of patients ≥18 years old in the ED with CT-documented SDH from 2014-2016. Patients were considered high risk per the decision instrument if they had hematoma thickness >5mm, warfarin or clopidogrel use, initial Glasgow Coma Scale <14, midline shift, or ≥1 hematoma. Primary composite outcome included in-hospital death or transition to hospice, need for neurosurgery, neurologic decline, and/or worsening hematoma on CT. Five hundred twenty-three charts were reviewed, and 222 met inclusion criteria. Forty-one (18.5%) patients were low risk per the decision instrument. Of these, 13 (31.7%) experienced the composite outcome. The sensitivity for this instrument was 89.4% (95% CI 82.6% - 94.3%), and the specificity was 28.3% (95% CI 19.7% - 38.2%). Patients on novel oral anticoagulant (NOAC) agents had a 55.6% (5/9) composite outcome. Including all patients on NOACs as high risk did not affect sensitivity (89.4%, 95% CI 82.6% - 94.3%) and marginally worsened specificity (24.2%, 95% CI 16.2% - 33.9%). The current decision instrument resulted in a 31.7% adverse outcome rate for low risk patients. This could be due to differences in our patient population including higher rate of NOAC use, inclusion of patients with spontaneous SDH, and higher overall mortality in our population. We recommend further study before implementing reduced level of care for patients with SDH.