Differences in glenohumeral translations calculated with three methods: Comparison of relative positions and contact point

Several published articles have reported 3-dimensional glenohumeral kinematics using model-image registration techniques. However, different methods to compute the translations were used in these articles. The purpose of this study was to compare glenohumeral translations calculated with three different methods. Fifteen healthy males with a mean age of 31 years (range, 27-36 years old) were enrolled in this study. Fluoroscopic images during scapular plane elevation were recorded at 30 frames per second for the right shoulder in each subject, and CT-derived models of the humerus and the scapula were matched with the silhouette of the bones in the fluoroscopic images using model-image registration techniques. Glenohumeral translations were computed with three methods: relative position of the origins of the humeral and scapular models, contact points of the two models, and relative positions based upon the calculated glenohumeral center of rotation (CoR). In the supero-inferior direction, translations calculated with the three methods were roughly parallel, with the maximum difference of 1.6 mm (P<0.001). In the antero-posterior direction, translations with the origins and CoR were parallel; however, translations computed with the origins and contact point describe arcs that differ by almost 2 mm at low humeral elevation angles and converge at higher degrees of humeral elevation (P<0.001). Glenohumeral translations calculated using three methods showed statistically significant differences that may be important when comparing detailed results of different studies. However, these relatively small differences are likely subclinical, so that all three methods can reasonably be used for description of glenohumeral translations.