Decompressive craniectomy or medical management for refractory intracranial hypertension: An AAST-MIT propensity score analysis

Background: Moderate/severe traumatic brain injury (TBI) management involves minimizing cerebral edema to maintain brain oxygen delivery. While medical therapy (MT) consisting of diuresis, hyperosmolar therapy, ventriculostomy, and barbiturate coma is the standard of care, decompressive craniectomy (DC) for refractory intracranial hypertension (ICH) has gained renewed interest. Since TBI treatment guidelines consider DC a second-tier intervention after MT failure, we sought to determine if early DC (<48 hours) was associated with improved survival in patients with refractory ICH. Methods: Eleven Level 1 trauma centers provided clinical data and head computed tomographic scans for patients with a Glasgow Coma Scale (GCS) score of 13 or less and radiographic evidence of TBI excluding deaths within 48 hours. Computed tomographic scans were graded according to the Marshall classification. A propensity score to receive DC (regardless of whether DC was performed) was calculated for each patient based on patient characteristics, physiology, injury severity, GCS, severity of intracranial injury, and treatment center. Patients who actually received a DC were matched to patients with similar propensity scores who received MT for analysis. Outcomes were compared between early (<48 hours of injury) primary or secondary DC and matched controls and then between early primary DC only and matched controls. Results: There were 2,602 patients who met the inclusion criteria ,of whom 264 (10.1%) received DC (either primary or secondary to another cranial procedure) and 109 (5%) had a DC that was primary. Variables associated with performing a DC included sex, race, intracranial pressure monitor placement, in-house trauma attending, traumatic subarachnoid hemorrhage, midline shift, and basal cistern compression. There was no survival benefit with early primary DC compared with the controls (relative risk, 1.07; 95% confidence interval, 0.67-1.73; p = 0.77), and resource use was higher. Conclusion: Early DC does not seem to significantly improve mortality in patients with refractory ICH compared with MT. Neurosurgeons should pause before entertaining this resource-demanding form of therapy. LEVEL OF EVIDENCE: Therapeutic care/management, level III.