The optimal screening and treatment algorithms for blunt cerebrovascular injury (BCVI) in cervical spine fracture patients remain unknown. Designing screening protocols for rare but potentially devastating conditions is difficult due to the risk of both false positives leading to overtreatment as well as false negatives leading to missed diagnosis. Rare conditions (i.e. BCVI leading to stroke) are very difficult to study due to lack of numbers, so there are no high quality prospective studies evaluating screening and treatment algorithms for BCVI in cervical spine fracture patients. Most of the treatment recommendations (i.e. antithrombotic or antiplatelet therapy for patients with asymptomatic, low grade BCVI) are based on retrospective studies confounded by selection bias as the untreated group typically had more severe injuries that precluded antithrombotic or antiplatelet therapy and thus likely sustained higher energy trauma that is associated with a higher risk of stroke. The current protocols used at most Level I trauma centers involve CT angiorgraphy (CTA) screening all patients with C1-C3 fractures, even low energy odontoid fractures or nondisplaced fractures of the C1 ring not involving the foramen transversarium. This leads to high screening rates in patients at low risk for BCVI associated stroke, which results in the identification and treatment of low grade BCVI. The stroke rate in these patients is exceedingly low, and anticoagulation in this population can be associated with a substantial risk of bleeding complications. The trauma community is in need of a more specific screening algorithm in order to prevent false positives and potentially harmful overtreatment. In order to address this, Dr. Fourman and colleagues from the University of Pittsburgh developed a screening algorithm that includes both trauma (i.e. high energy trauma, loss of consciousness, or altered mental status) and fracture (transverse foramen involvement, combined C1-C2 fracture, bilateral facet fracture, or facet dislocation) charcteristics. In a retrospective cervical spine fracture cohort of 721 patients, use of this algorithm would have resulted in CTA being performed in 36% of patients compared to about 57% of patients using traditional screening protocols. The new protocol would have missed 16 vertebral artery injuries detected by the other protocols but would have missed none of the 7 patients who were diagnosed with BCVI associated stroke. One of the stroke patients was missed by the traditional screening algorithm. Based on these findings, the authors suggested that their new algorithm could reduce the rate of unnecessary CTA without missing BCVI associated stroke.
The authors have done a nice job attempting to create an algorithm that reduces the need for CTA, which is being performed with increasing frequency in trauma patients. This test not only costs money and exposes patients to radiation, but it also identifies many low grade BCVIs, for which the benefit of antithrombotic or antiplatelet therapy is unclear. The main limitation of this paper is that most of the BCVI patients were likely treated, and treatment may have prevented stroke. As such, while the proposed algorithm would have resulted in CTA and identification of BCVI in all stroke patients, it is possible that some of the treated BCVI patients who did not have a stroke would have been missed by the current algorithm, would not have been treated, and would have had a stroke as a result. It is not possible to control for this limitation with a retrospective study. The authors also did not address how they identified the criteria used in their algorithm, and it is possible that other criteria would lead to a more accurate algorithm. While there is a definite need for better algorithms for BCVI screening and treatment, retrospective studies have significant limitations and will probably not accomplish this. This and other similar papers should provide an impetus for a large, multicenter study looking at this topic. However, such a study would be difficult as it would either need to be randomized (i.e. CTA or no CTA, treatment or no treatment of identified low grade BCVI) or involve screening a large number of patients and then not treating low grade BCVI (possibly ethically unsound).
Please read Dr. Fourman's article on this topic in the December 1 issue. Does this change how you view BCVI screening algorithms? Let us know by leaving a comment on The Spine Blog.
Adam Pearson, MD, MS
Associate Web Editor