Background: Acute management of traumatic brain injury (TBI), in particular mild TBI, focuses on the detection of the 5-7 % who may be harboring potentially life-threatening intracranial hemorrhage (IH) using CT scanning. Guidelines intending to reduce unnecessary head CT scans using available clinical variables to detect those at high IH risk have shown varying results. Recently, the Scandinavian Neurotrauma Committee (SNC) derived a new set of high-IH risk variables for adults with TBI using an evidence-based literature review. Unlike previous guidelines, the SNC guideline incorporates serum values of the brain protein S100B with clinical variables.
Methods: We performed a nested cohort study of adults with mild TBI presenting to six emergency departments in New York and Pennsylvania within 6 h of injury. Patients were managed according to existing guidelines for CT selection. All patients underwent head CT scanning and serum S100B measurement, as well as prospective collection of clinical variables, as a requirement of the parent study. Using the SNC guidelines, S100B values and clinical variables were applied to these subjects, classifying each into one of five pre-defined severity categories, as well as predicting the need for head CT scanning to identify IH. This classification was then compared to actual head CT results to determine guideline sensitivity and specificity.
Results: In total, 662 adults (mean age 42 years, range 18-96; 258 females, 549 Caucasians) were available for analysis; 36 (5%) had IH on head CT scan. The SNC guidelines had a sensitivity of 97% (95% CI, 84-100%) and a specificity of 34% (95% CI, 30-37%) for the detection of IH on head CT. Application of the SNC guidelines would have resulted in a CT reduction of 32% (211/662 patients). One patient with low-risk mild TBI and a S100B level under 0.10 μg/L had a traumatic CT abnormality and would have been discharged with strict adherence to the guidelines. However, this patient did not need any intervention for the injury and had a good outcome.
Conclusion: Using the SNC guideline could save approximately one third of CT scans in a pre-selected cohort of mild TBI patients with little or no impact on patient outcome.