Adverse Outcomes After Initial Non-surgical Management of Subdural Hematoma: A Population-Based Study

Background: Little is known about the natural history of non-surgically managed subdural hematoma (SDH). The purpose of this study is to determine rates of adverse events after non-surgical management of SDH and whether these outcomes differ depending on traumatic versus nontraumatic etiology. A retrospective cohort study was conducted using administrative claims data on all emergency department visits and acute care hospitalizations at nonfederal facilities in California from 2005 to 2011, Florida from 2005 to 2012, and New York from 2006 to 2011. We included patients who were discharged home after hospitalization with a first-recorded diagnosis of SDH and no record of surgical hematoma evacuation.

Methods: Patients were followed for readmission with SDH, readmission for surgical SDH evacuation, and fatal readmission with SDH. Survival statistics and the log-rank test were used to compare rates of these adverse events after traumatic versus nontraumatic SDH. Multivariable Cox regression analysis was used to compare hazards for traumatic versus nontraumatic etiology while adjusting for age, sex, race, insurance status, presence of dementia, alcohol use, acquired abnormalities in coagulation, acquired abnormalities in platelet function, hypertension, atrial fibrillation, venous thromboembolism, ischemic stroke, coronary heart disease, and valvular disease.

Results: We identified 27,502 conservatively treated patients with SDH, of which 70.9% were traumatic and 29.1% nontraumatic. Compared to patients with traumatic SDH, patients with nontraumatic SDH had significantly higher rates of subsequent hospitalization with SDH (cumulative 90-day rates: 15.3 % [95% CI 14.5-16.1%] vs. 10.3% [95% CI 9.9-10.8%]), surgical SDH evacuation (7.8% [95% CI 7.3-8.5%] vs. 5.5% [95% CI 5.2-5.8%]), and SDH-related in-hospital death (1.0% [95% CI 0.8-1.2%] vs. 0.4% [95 % CI 0.3-0.5%]). In multivariable Cox regression analysis, nontraumatic etiology was associated with a higher hazard of readmission with SDH (HR 1.4; 95% CI 1.3-1.5), surgery (HR 1.3; 95% CI 1.2-1.4), and in-hospital mortality (HR 1.9; 95% CI 1.4-2.5). Our findings were unchanged in sensitivity analyses that also adjusted for Elixhauser comorbidities.

Conclusions: Approximately one in eight patients with a conservatively managed SDH was readmitted with SDH within 90 days. A substantial proportion of these readmissions involved surgical hematoma evacuation. These outcomes occurred significantly more often after nontraumatic as compared to traumatic SDH.