Standard radiological classification of glenohumeral osteoarthritis does not correlate with the complexity of the arthritic glenoid deformity.

Background: The Samilson-Prieto classification (SPC) depending on the humeral osteophyte length on a-pX-rays today is widely used to classify glenohumeral osteoarthritis in general. For treatment planning and prognosis, the patho-morphology of the glenoid and static posterior subluxation of the humeral head classified according to Walch is of much higher importance. Here, usually a CT or MRI scan is required for a correct classification. A possible correlation between both classifications is poorly explored. Without it, the complexity of the case might be mis-interpreted using the SPC. The aim of this study was to investigate such a correlation, i.e. whether it correlates with the glenoid deformity and degree of humeral head subluxation. Patients and methods: Radiological datasets (X-ray and CT or MRI) of 352 patients with primary OA of the shoulder were evaluated by two observers experienced in shoulder surgery. For the Samilson-Prieto classification, true a-p shoulder radiographs and for the modified Walch classification CT or MRI scans in the axial plane were interpreted using a validated method. To investigate a correlation between both classifications, the Fisher's exact test was used. For the interobserver reliability, the weighted kappa coefficient κ was determined. Results: For the Walch classification, both observers found a similar percentage for the different types, with decreasing numbers from normal (type A1) to severely altered glenoids In the Samilson-Prieto classification, OA grade I was predominant, while grade II and III showed a relatively equal distribution. Interobserver reliability was high both for the Walch classification with a κ 0.923 (95% confidence interval 0.892; 0.954) and) for the SPC with a κ 0.88 (95% confidence interval 0.843; 0.916). A correlation between the two classifications in Fischer's exact test could not be shown (p = 0.584). Discussion: Since there is no correlation between both, using the Samilson-Prieto classification alone might miss relevant prognostic factors in gleno-humeral OA. Adequate imaging of the glenoid morphology also in the axial plane is absolutely mandatory to understand the complexity and chose the right treatment for each patient. Level of evidence: Study of Diagnostic Test—Level II. [ABSTRACT FROM AUTHOR]