Predictors of Hospital Mortality in Older Patients with Subdural Hematoma

OBJECTIVE: To identify presenting characteristics of older patients with subdural hematoma who are unlikely to survive hospitalization. DESIGN: A retrospective cohort. PATIENTS: All patients > or = 65 years of age hospitalized at a tertiary care center from 1985-1990 with the primary diagnosis of subdural hematoma. MEASUREMENTS: Chart review was performed to characterize presenting clinical characteristics, hospital course, and outcome at the time of hospital discharge. MAIN RESULTS: Of the 157 eligible patients, 42% (66/157) were > or = 80 years of age. Although 30% of patients had no recorded trauma, 54% experienced a fall before hospitalization. Twenty-six percent (47/157) of patients had been on antithrombotic therapy (14 on coumadin, two on heparin, 31 on antiplatelet agents, one on both coumadin and an antiplatelet agent). Sixty percent of patients had no focal neurologic findings, and the mean Glasgow Coma Score was 12.3 (+/- 3.6). The hematoma was considered chronic in 49% (77/157) of cases, based on time from antecedent trauma or on neuroimaging criteria. Midline shift by neuroimaging was present in 69%. Overall hospital mortality was 31% (48/157). Using logistic regression to control for other factors, level of consciousness (Glasgow Coma Score or = 80 (OR = 3.7), duration of hematoma considered acute (OR = 2.7), and craniotomy (OR = 2.6) were significantly associated with hospital mortality. Presence of focal symptoms, previous antithrombotic medication use, nature of trauma, comorbidity score, and presence of midline shift were not associated with hospital mortality. CONCLUSIONS: Among older patients with subdural hematoma, level of consciousness, extreme old age, duration of the hematoma, and nature of the intervention were significantly associated with hospital mortality. These factors should help physicians in clinical decision making and formulation of advance directives for geriatric patients with subdural hematoma.