Blunt cerebrovascular injury (BCVI) can result in thromboembolic stroke. Many trauma centers selectively screen patients with cervical computed tomographic angiography (CTA) based on clinical criteria. In 2016, our institution adopted universal screening for BCVI for all blunt trauma patients. The aim of this study was to accurately determine the incidence of BCVI and to evaluate the diagnostic performance of the Denver criteria (DC), expanded Denver criteria (eDC), and Memphis criteria (MC) in selecting patients for screening.
Retrospective cohort study of adult (≥16 years) blunt trauma patients who presented to the Level I trauma center at University of Alabama at Birmingham. We reviewed all CTA reports and selected CTA images to obtain the true incidence rate of BCVI. We then evaluated the diagnostic performance of the DC, eDC, and MC.
A total of 6,800 patients who had suffered blunt trauma were evaluated, of whom 6,287 (92.5%) had a neck CTA. Of these, 480 (7.6%) patients had CTA evidence of BCVI. The eDC identified the most BCVI cases (sensitivity 74.7%) but had the lowest positive predictive value (14.6%). The DC and MC had slightly greater positive predictive values (19.6% and 20.6%, respectively) and had the highest diagnostic ability in terms of likelihood ratio (2.8 and 2.9) but had low sensitivity (57.5% and 47.3%). Consequently, if relying on the traditional screening criteria, the DC, eDC, and MC would have respectively resulted in 42.5%, 25.3%, and 52.7% of patients with BCVI identified by universal screening not receiving a neck CTA to screen for BCVI.
Blunt cerebrovascular injury is even more common than previously thought. The diagnostic performance of selective clinical screening criteria is poor. Consideration should be given to the implementation of universal screening for BCVI using neck CTA in all blunt trauma patients.
LEVEL OF EVIDENCE
Diagnostic, level III.