Surgical management of traumatic brain injury: a comparative-effectiveness study of 2 centers

Mass lesions from traumatic brain injury (TBI) often require surgical evacuation as a life-saving measure and to improve outcomes, but optimal timing and surgical technique, including decompressive craniectomy, have not been fully defined. The authors compared neurosurgical approaches in the treatment of TBI at 2 academic medical centers to document variations in real-world practice and evaluate the efficacies of different approaches on postsurgical course and long-term outcome.

Patients 18 years of age or older who required neurosurgical lesion evacuation or decompression for TBI were enrolled in the Co-Operative Studies on Brain Injury Depolarizations (COSBID) at King's College Hospital (KCH, n = 27) and Virginia Commonwealth University (VCU, n = 24) from July 2004 to March 2010. Subdural electrode strips were placed at the time of surgery for subsequent electrocorticographic monitoring of spreading depolarizations; injury characteristics, physiological monitoring data, and 6-month outcomes were collected prospectively. CT scans and medical records were reviewed retrospectively to determine lesion characteristics, surgical indications, and procedures performed.

Patients enrolled at KCH were significantly older than those enrolled at VCU (48 vs 34 years, p < 0.01) and falls were more commonly the cause of TBI in the KCH group than in the VCU group. Otherwise, KCH and VCU patients had similar prognoses, lesion types (subdural hematomas: 30%-35%; parenchymal contusions: 48%-52%), signs of mass effect (midline shift ≥ 5 mm: 43%-52%), and preoperative intracranial pressure (ICP). At VCU, however, surgeries were performed earlier (median 0.51 vs 0.83 days posttrauma, p < 0.05), bone flaps were larger (mean 82 vs 53 cm(2), p < 0.001), and craniectomies were more common (performed in 75% vs 44% of cases, p < 0.05). Postoperatively, maximum ICP values were lower at VCU (mean 22.5 vs 31.4 mm Hg, p < 0.01). Differences in incidence of spreading depolarizations (KCH: 63%, VCU: 42%, p = 0.13) and poor outcomes (KCH: 54%, VCU: 33%, p = 0.14) were not significant. In a subgroup analysis of only those patients who underwent early (< 24 hours) lesion evacuation (KCH: n = 14; VCU: n = 16), however, VCU patients fared significantly better. In the VCU patients, bone flaps were larger (mean 85 vs 48 cm(2) at KCH, p < 0.001), spreading depolarizations were less common (31% vs 86% at KCH, p < 0.01), postoperative ICP values were lower (mean: 20.8 vs 30.2 mm Hg at KCH, p < 0.05), and good outcomes were more common (69% vs 29% at KCH, p < 0.05). Spreading depolarizations were the only significant predictor of outcome in multivariate analysis.

This comparative-effectiveness study provides evidence for major practice variation in surgical management of severe TBI. Although ages differed between the 2 cohorts, the results suggest that a more aggressive approach, including earlier surgery, larger craniotomy, and removal of bone flap, may reduce ICP, prevent cortical spreading depolarizations, and improve outcomes. In particular, patients requiring evacuation of subdural hematomas and contusions may benefit from decompressive craniectomy in conjunction with lesion evacuation, even when elevated ICP is not a factor in the decision to perform surgery.