Shoulder arthroplasty, both primary (TSA) and reverse (RTSA), are common interventions for arthritis and cuff tear arthropathy. The effect of shoulder arthroplasty on shoulder motion is of particular interest in assessing the effectiveness of the procedure and the development and biomechanical testing of implants. A comparison of the arthroplasty shoulder to that of the non-operated contralateral shoulder provides insight into how well the reconstruction has restored natural shoulder motion. The purpose of this study was to ascertain the shoulder motion of patients who have undergone shoulder arthroplasty and to compare the motion of the reconstructed and contralateral natural sides.

Eleven human subjects (70±9yrs) who had undergone total shoulder arthroplasty wore a custom instrumented shirt for the waking hours of one day. The 3D orientation of each humeral sensor was transformed with respect to the torso to allow for the calculation of humeral elevation and plane of elevation angles. Joint angles for each subject were then discretised, and the operative and contralateral normal (control) shoulders were then compared.

The majority of both the arthroplasty and control shoulder elevation motions took place below 80° of elevation, totaling on average 1910±373 and 1887±312 motions per hour, respectively. Conversely, elevations greater than 80° were significantly less with occurrences totaling only 55±31 and 78±41 motions per hour for the arthroplasty and control shoulders, respectively (p<0.01). Both the arthroplasty and control shoulder were at elevations below 80° for 88±7% and 87±7% of the day, respectively. When the total motion of the arthroplasty and non-operative control shoulders were compared, no statistically significant difference was detected (p=0.8), although the non-operated side exhibited marginally more motion than the operated side, an effect which was larger at higher elevation angles (p=0.3).

This study provides insight into the effects of shoulder arthroplasty on thoraco-humeral motion and compares it to the non-operative side. Interestingly, there were no significant differences measured between the arthroplasty and the control side, which may demonstrate the effectiveness of reconstruction on restoring natural shoulder motion. It is interesting to note that on average, each shoulder arthroplasty elevated above 80° approximately 55 times per hour, corresponding to just under 330,000 motions per year. Similarly, when elevations greater than 60° are extrapolated, the resulting yearly motions total approximately 1.5 million cycles (Mc), which suggests that the ‘duty cycle’ of the shoulder is similar to the hip, approximated to be between 1–2 Mc per year. Arthroplasty wear simulators should be calibrated to simulate these patterns of motion, and component design may be improved by understanding the kinematics of actual shoulder motion.