The critical shoulder angle (CSA) in glenohumeral osteoarthritis: Does observer experience affect measurement reliability on plain radiographs?

Abstract

The critical shoulder angle (CSA) has been associated with full-thickness rotator cuff tears both in the presence and absence of glenohumeral arthritis. It is unclear whether the CSA can be reliably measured from plain radiographs of concentric glenohumeral osteoarthritis amongst examiners at differing levels of training. We retrospectively reviewed the radiographs of consecutive patients who underwent shoulder arthroplasty for glenohumeral osteoarthritis. The CSA was measured on a standardized AP scapular view at baseline and then 4 weeks later by fellowship-trained orthopaedic surgeons, a shoulder fellow and a senior orthopaedic resident. Grade of arthritis was categorized using the Samilson and Prieto method. The inter- and intra-observer reliability was then determined for all examiners, as well as for increasing severity of radiographic arthritis. The relationship between the CSA and grade of arthritis was assessed. There were 166 included patients comprised of 104 females (63%) and 62 males (37%) with a mean age of 65.9 ± 10.4 years. The inter- and intra-observer reliability for measuring the CSA amongst all examiners was found to be excellent, with an intra-class coefficient (ICC) of >0.9 (p < 0.0001). The ICC remained excellent even amongst radiographs with more advanced arthritis. Furthermore, there was a weak, inverse relationship between the grade of arthritis and the CSA (r = -0.377, p < 0.005). The CSA can reliably be measured by examiners at varying levels of orthopaedic training, even with more advanced radiographic glenohumeral osteoarthritis. Level of evidence: Level III (Prognostic).