A new classification of glenoid bone loss to help plan the implantation of a glenoid component before revision arthroplasty of the shoulder.

Aims Glenoid bone loss can be a challenging problem when revising a shoulder arthroplasty. Precise pre-operative planning based on plain radiographs or CT scans is essential. We have investigated a new radiological classification system to describe the degree of medialisation of the bony glenoid and that will indicate the amount of bone potentially available for supporting a glenoid component. It depends on the relationship between the most medial part of the articular surface of the glenoid with the base of the coracoid process and the spinoglenoid notch: it classifies the degree of bone loss into three types. It also attempts to predict the type of glenoid reconstruction that may be possible (impaction bone grafting, structural grafting or simple non-augmented arthroplasty) and gives guidance about whether a pre-operative CT scan is indicated. Patients and Methods Inter-method reliability between plain radiographs and CT scans was assessed retrospectively by three independent observers using data from 39 randomly selected patients. Inter-observer reliability and test-retest reliability was tested on the same cohort using Cohen's kappa statistics. Correlation of the type of glenoid with the Constant score and its pain component was analysed using the Kruskal-Wallis method on data from 128 patients. Anatomical studies of the scapula were reviewed to explain the findings. Results Excellent inter-method reliability, inter-observer and test-retest reliability were seen. The system did not correlate with the Constant score, but correlated well with its pain component. Take home message: Our system of classification is a helpful guide to the degree of glenoid bone loss when embarking on revision shoulder arthroplasty.