A Decision Instrument to Identify Isolated Traumatic Subdural Hematomas at Low Risk of Neurologic Deterioration, Surgical Intervention, or Radiographic Worsening

Abstract

Objectives: Subdural hematoma (SDH) is the most common form of traumatic intracranial hemorrhage. Severity of disease in patients with SDH varies widely. It was hypothesized that a decision rule could identify patients with SDH who are at very low risk for neurologic decline, neurosurgical intervention, or radiographic worsening. Methods: Retrospective chart review of consecutive patients age ≥ 16 with Glasgow Coma Score (GCS) ≥ 13 and computed tomography (CT)-documented isolated SDH presenting to a university-affiliated, urban, 100,000-annual-visit ED from 2009 to 2015. Demographic, historical, and physical examination variables were collected. Primary outcome was a composite of neurosurgical intervention, worsening repeat CT, and neurologic decline. Univariate analysis was performed and statistically important variables were utilized to create a logistic regression model. Results: A total of 644 patients with isolated SDH were reviewed, 340 in the derivation group and 304 in the validation set. Mortality was 2.2%. A total 15.5% of patients required neurosurgery. A decision instrument was created: patients were low risk if they had none of the following factors—SDH thickness ≥ 5mm, warfarin use, clopidogrel use, GCS < 14, and presence of midline shift. This model had a sensitivity of 98.6% for the composite endpoint, specificity of 37.1%, and a negative likelihood ratio of 0.037. In the validation cohort, sensitivity was 96.3%, specificity was 31.5%, and negative likelihood ratio was 0.127. Conclusion: Subdural hematomas are amenable to risk stratification analysis. With prospective validation, this decision instrument may aid in triaging these patients, including reducing the need for transfer to tertiary centers.