Progressive Displacement of Scapula Fractures
Journal of Trauma
INTRODUCTION Historically, minimally-to-moderately displaced extra-articular scapula fractures have been managed conservatively. It is possible that operative criteria could be met for a given patient if progressive displacement in the post injury period were to occur. A retrospective review of an operative cohort of patients was performed to determine if a subset of scapula fractures that initially were minimally displaced and did not meet operative criteria subsequently showed progressive
... wed progressive displaced in the post-injury period, thereby meeting operative criteria. METHODS This is a retrospective review of 49 patients with displaced, extraarticular scapula neck fractures who underwent open reduction and internal fixation for a scapula fracture between 2003 and 2007. Indications for operative treatment included at least one of the following criteria: >20 mm of medialization of the glenohumeral joint, >25 degrees of angular deformity in the semicoronal plane as seen in the scapula Y view, or displaced (>10 mm) double lesions of the superior shoulder suspensory complex. The patients were screened to select out those who were initially managed conservatively based on the initial clinical and radiographic assessment, but then subsequently were found to meet operative indications due to progressive displacement/deformity noted on evaluation of repeat radiography at their next follow-up visit. Eight out of 49 patients (16%) who underwent surgical treatment were initially managed conservatively and then subsequently met surgical criteria due to progressive deformity noted at follow-up. Radiographic measurements for deformity, including medialization, angulation, translation, and glenopolar angle (GPA), taken at the initial, post-injury assessment were then compared to those taken at the followup assessment, and the difference was calculated for each of these deformities. Medialization in this context is defined as displacement of the proximal fragment (attached to the glenohumeral joint) in relation to the distal fragment (scapula body) resulting in lateral border offset. To examine inter-observer variability, three physician examiners of different levels of experience were assigned to independently measure radiographs. These included a board-certified orthopaedic surgeon specializing in orthopaedic trauma, an orthopaedic trauma fellow, and a clinical research fellow. Each observer was blinded to the others' readings. Intraobserver reliability was evaluated by having each of the examiners measure the same set of radiographs on two separate occasions, 16 weeks apart. Subsequently, single-measure intraclass correlation coefficients (ICCs) were used to determine variability within and among measurement groups . According to Shrout and Fleiss  an ICC of 0.6-0.74 is considered good and an ICC of 0.75 is excellent. Pearson's r coefficients were also calculated to determine intra-observer variability. Statistical analyses consisted of the mean, range, standard deviation, and t-test. For all analyses, statistical significance was assumed for p < 0.01. All the imaging available including radiographs and 3D-CT scans were evaluated and those views that allowed best visualization of the fracture line were collected for each patient. These views from each patient were imported into Macromedia Fireworks MX software (Macromedia Inc., San Francisco, CA) to overlap and orient fracture patterns onto a template scapula image. Images of each scapula were graphically superimposed to create a compilation of fracture lines on a standard scapula template. The overlap of all fracture lines resulted in a frequency diagram. Once proper anatomical alignment was obtained, fracture lines were identified and were traced on top of the combined 3D-CT and model scapula. Fracture patterns were then confirmed using the original 3D-CT rendering. RESULTS Eight out of 49 patients (16%) who underwent surgical treatment were initially managed conservatively and then subsequently met surgical criteria due to progressive deformity noted at follow-up. Three patients were female and 5 patients were male with an average age of 38 (range 19-53). All fractures were a result of high-energy trauma. Overlaying of the fracture patterns yielded an illustration of a fracture map for this cohort (Fig.