Contemporary outcomes of vertebral artery injury

Objective: Vertebral artery injury (VAI) associated with cervical trauma is being increasingly recognized with more aggressive screening. Disparate results from previous literature have led to uncertainty of the significance, natural history, and optimal therapy for VAI. Methods: To understand the natural history and treatment outcomes from our experience, we performed a retrospective, single-center review from a level I trauma center for the previous 10 years of all VAI. Injuries were identified from search of an administrative trauma database, a resident-run working database, and all radiology dictations for the same period. All VAI were classified according to segmental involvement, Denver grading scale, and laterality. Analysis of associated injuries, demographics, neurologic outcome, mortality, length of stay, treatment plan, and follow-up imaging was also performed. Results: Fifty-one patients with VAI were identified from 2001 to 2011 from a total of 36,942 trauma admissions (0.13% incidence). Associated injuries were significant with an average New Injury Severity Score of 29.6. Penetrating trauma occurred in 14%. Cervical spine fracture was present in 88% with VAI. Diagnosis was obtained with computed tomographic angiography (CTA) in 95%. Screening was prompted by injury pattern or high-risk mechanism in all cases. Injuries classified according to the Denver grading scale were grade I = 24%, grade II = 35%, grade III = 4%, grade IV = 35%, and grade V = 2%. Distribution across segments included V1 = 18%, V2 = 67%, V3 = 31%, and V4 = 6%. Only one posterior circulation stroke was attributable to VAI. Overall mortality was 8%, with each mortality being associated with significant other organ injuries. Treatment rendered for VAI was antiplatelet therapy (50%), observation (31%), warfarin (17%), and stent (2%). There were no significant differences between treatment groups on any variable with the exception of body mass index (P = .047). Follow-up was obtained for 13% (n = 6) of survivors. The CTA demonstrated injury stability in four patients and resolution in two patients. Accuracy of the administrative trauma database was 53% compared with 96% for the resident-run working database. Conclusions: Neurologic sequelae attributable to VAI were rare. Grade of VAI or vertebral artery segment did not correlate with morbidity. We did not observe any differences in short-term outcomes between systemic anticoagulation and antiplatelet therapy. Of those patients seen at follow-up, injury resolution or stability was documented by CTA. A conservative approach with either observation or antithrombotic therapy is suggested. If the natural history of VAI includes a very low stroke rate, then therapies with a lower therapeutic index, such as systemic anticoagulation, in the severely injured trauma patient are not supported. Our search strategy urges awareness of the limitations of administrative databases for retrospective vascular study. (J Vasc Surg 2013;57:741-6.)