Operative neurosurgery for traumatic subdural hematoma: association between trauma center variation and patient outcomes

OBJECTIVE: Traumatic subdural hematoma (SDH) is a common form of traumatic brain injury (TBI) that often represents a neurosurgical emergency. Surgical evacuation is recommended for SDH with midline shift (MLS) > 5 mm, regardless of the presenting Glasgow Coma Scale (GCS) score; however, real-world practice is unknown. The objective of this study was to test the hypothesis that significant variation exists in the tendency for operative neurosurgical intervention for traumatic SDH among trauma centers (TCs) and that this variation is associated with patient outcomes. METHODS: The authors performed a retrospective cohort study of adult patients (age ≥ 18 years) treated for blunt severe TBI (GCS score ≤ 8) and SDH with MLS > 5 mm at level I and II TCs participating in the American College of Surgeons Trauma Quality Improvement Program (2017-2019). Patients with nonsurvivable injuries (Abbreviated Injury Scale score 6), advance directives, or emergency department death were excluded. Hierarchical logistic regression was used to estimate each TC's unique odds of performing operative neurosurgery for traumatic SDH. Risk adjustment accounted for patient baseline and injury characteristics. TCs were grouped into quartiles of increasing tendency for neurosurgery. The risk-adjusted association between TC tendency for operative neurosurgery and outcomes was then measured. The primary outcome was inpatient mortality. The secondary outcome was favorable discharge, defined as discharge to home or rehabilitation. RESULTS: The authors identified 13,087 patients with traumatic SDH treated at 454 level I and II TCs. Significant variation in TC tendency for operative neurosurgery was observed. Specifically, TCs with the greatest tendency for neurosurgical intervention (quartile 4) performed surgery on 60% of patients, whereas TCs with the lowest tendency (quartile 1) performed surgery on only 26%, even though there were no differences in GCS scores or pupillary examination findings. After risk adjustment, a greater hospital tendency for neurosurgical intervention was associated with lower inpatient mortality and higher odds of favorable discharge. Patients with traumatic SDH treated at TCs with the highest versus the lowest tendency for neurosurgery were 30% less likely to die (adjusted odds ratio [aOR] 0.7, 95% CI 0.6-0.8) and more likely to have a favorable discharge (aOR 1.3, 95% CI 1.1-1.6). These effects were most pronounced among patients with abnormal pupillary examination findings. CONCLUSIONS: Significant variation exists between trauma centers in performing operative neurosurgery for traumatic SDH. The TCs more likely to perform surgery were associated with lower odds of inpatient mortality and higher odds of a favorable discharge. Consensus-based guidelines are needed to improve standardization in the care of patients with traumatic SDH.