Background Scapula fractures are rare injuries ( 2 cm, glenopolar angle ≤ 22°, angulation ≥ 45°, double disruption of the shoulder suspensory complex, or open fractures), what was the median lateral border displacement beyond the edge of the glenoid (spike sign)? (2) What amount of displacement (spike sign) indicates surgery based on previously established criteria, and what are the sensitivity, specificity, positive predictive value, and negative predictive value of the displacement threshold for indicating surgery? (3) What are the interobserver and intraobserver reliability of the spike sign? Methods This was a retrospective, diagnostic study evaluating the diagnostic performance of a radiographic finding—the spike sign—as a marker for fractures that would meet reasonable, widely accepted surgical indications. The work was performed at one large, urban, Level I trauma center. Patients with AO Foundation/Orthopaedic Trauma Association (AO/OTA) 14B extraarticular fractures of the scapula from 2013 to 2021 were identified. Utilizing a comprehensive search of the electronic medical record, patients who were diagnosed with a scapula fracture were identified. Between September 2013 and December 2021, a total of 400 patients at one academic Level I trauma center were noted to have a diagnosis of scapular fractures. Thirty-three percent (132 of 400) were excluded because of missing data, inadequate or missing radiographs, open fractures, intraarticular fractures, isolated process fractures, malunion, nonunion, or patient death from initial trauma, leaving 67% (268 of 400) for analysis, the focus of this retrospective study. Of those, 50% (133 of 268) met one or more surgical indications: medialization (lateral border offset) ≥ 2 cm, angulation ≥ 45°, glenopolar angle ≤ 22°, or double disruption of the shoulder suspensory complex. The mean ± SD patient age was 50 ± 17 years. Twenty-one percent (55 of 268) of the group was female and 81% (217) were White. The most common mechanism of injury was motorcycle collision, occurring 26% (70) of the time. Twenty-nine percent (77) of patients had measurable lateral border displacement. The spike sign is the caudal lateral border of the fractured scapula displacement beyond the glenoid on AP shoulder radiographs. We considered the following to be indications for surgical treatment of scapular fractures during the period in question, and meeting one or more of the following was the “gold standard” against which the spike sign was assessed: medialization (lateral border offset) ≥ 2 cm, angulation ≥ 45°, glenopolar angle ≤ 22°, or displaced double disruption of the shoulder with clavicle and scapular neck fracture. Open fractures and intraarticular fractures were excluded because they were not within the scope of this study. To answer our first question, we took measurements of the lateral border offset, medialization, and glenopolar angle of all 268 patients’ scapula injury radiographs. To answer our second question, we identified every scapula with an identifiable positive spike sign and evaluated whether those scapulae also met operative indications based on any of the previous measurements. To answer our final question, two orthopaedic trauma fellows took independent measurements for patients with identified lateral border displacement and were compared to determine interobserver and intraobserver reliability. Results In patients who meet surgical indications for extraarticular scapular fractures, the median lateral displacement with respect to the glenoid (spike sign) was -0.21 cm (range -2.48 cm to 2.66 cm). With the spike sign threshold set at ≥ 0 cm (spike at or lateral to the glenoid), the positive predictive value for surgery was 100% (95% confidence interval CI 91% to 100%). Sensitivity was 28% (95% CI 20% to 36%) and specificity was 100% (95% CI 97% to 100%). A spike sign ≥ 0 cm (spike lateral to the glenoid) was confirmed to be the appropriate threshold for a positive spike sign using receiver operating characteristic analysis. Fourteen percent (37 of 268) had a positive spike sign, all of whom met at least one operative indication (medialization, angulation, glenopolar angle, or double disruption of the shoulder). The positive predictive value for spike signs ≥ 0 cm was 100%. Two rounds of interobserver reliability for the spike sign calculated with intraclass correlation coefficient (0.87 and 0.94) and Cohen kappa (κ = 0.84 and 0.88) indicated excellent reliability and near-perfect agreement, respectively. Intraobserver reliability for Reviewer 1 was 93.75% agreement, and for Reviewer 2, it was 100%. Conclusion The scapular spike sign, when it occurs, is an easily identified radiographic finding on AP shoulder radiographs in extraarticular scapula fractures and may serve as a reliable proxy for surgical indications. In light of these findings, surgeons should be able to better identify complex scapula fractures in patients with a positive spike sign, which should prompt more efficient definitive management of these injuries even in settings where advanced imaging may not be readily available. Future studies should look to correlate the utility of the spike sign in the trauma setting and further evaluate whether this is now affecting the number of patients undergoing surgery to fix fractures that may have been managed nonoperatively in the past. Level of Evidence Level IV, diagnostic study.
Brahme et al. (Thu,) studied this question.