Blunt Thoracolumbar-Spine Trauma Evaluation in the Emergency Department: A Meta-Analysis of Diagnostic Accuracy for History, Physical Examination, and Imaging

Abstract Background Delayed diagnoses of unstable thoracolumbar spine (TL-spine) fractures can result in neurologic deficits and avoidable pain, so it is important for clinicians to reach prompt diagnostic decisions. There are no validated decision aids for determining which trauma patients warrant TL-spine imaging. Objective Our aim was to quantify the diagnostic accuracy of the injury mechanism, physical examination, associated injuries, clinical decision aids, and imaging for evaluating blunt TL-spine trauma patients. Methods A search strategy for studies including adult blunt TL-spine trauma using PubMed, Embase, Scopus, CENTRAL, Cochrane Database of Systematic Reviews, and ClinicalTrials.gov was performed. Excluded studies lacked data to construct 2 × 2 tables, were duplicates, were not primary research, did not focus on blunt trauma, examined associated injuries without any utility in identifying TL-spine injuries, only studied cervical-spine fractures, were non-English, had a pediatric setting, or were cadaver/autopsy reports. Risk of bias was assessed using the Quality Assessment Tool for Diagnostic Accuracy Studies. Diagnostic predictors were analyzed with a meta-analysis of sensitivity, specificity, and likelihood ratios. Results In blunt trauma patients in the emergency department, the weighted pretest probability of a TL-spine fracture was 15%. The estimates for detection of TL-spine fractures with plain film were: positive likelihood ratio (+LR) = 25.0 (95% confidence interval [CI] 4.1–152.2; I 2  = 94%; p 2  = 84%; p 2  = 87%; p 2  = 23%; p  = 0.26). Conclusions CT is more accurate than plain films for detecting TL-spine fractures. Injury mechanism, physical examination, and associated injuries alone are not accurate to rule-in or rule-out TL-spine fractures.