Reliability of proximal migration measurements in shoulder arthroplasty

Proximal migration is a complication in shoulder arthroplasty. Quantification of this phenomenon would help in identifying its cause. This study assessed the reliability of 4 methods of measuring proximal migration after shoulder arthroplasty on standard anteroposterior radiographs: the glenohumeral center-to-center measure, the glenohumeral arc measure, the acromiohumeral distance, and the scapular spine-humeral head center (spinohumeral) distance. Radiographs were taken of 12 embalmed shoulders, with matching metal humeral head replacements, in the neutral position and rotated both ways for 20 degrees along the vertical and horizontal axes. The measurements were tested for the reliability of the deviating projections against the neutral position, and an estimate of the interobserver and intraobserver reliability was made. Statistics included a paired t test and the interclass correlation coefficient. The acromiohumeral distance and spinohumeral center method proved most reliable. The spinohumeral center method was the least sensitive for projection errors, although only applicable as a relative measure. The coracoid process base can be used as a tell-tale sign for scapular projection.