33 patients with osteoarthritis of the shoulder and intact rotator cuffs were randomized to either hemiarthroplasty (HA) or total shoulder replacement (TSR). The aim of this study was originally to provide a short term comparison of these two treatments, but the trial was extended and the patients were progressively reviewed to a minimum of 10 year follow up to assess differences in pain, function, and revision rates. At 2 years, the TSR group reported less pain, and superior Constant and UCLA scores compared to the HA group. Trial enrollment was terminated in lieu of 2-year results, although follow-up was continued in those already entered into the study. At the 10 year follow up there were no differences in pain or function between the two groups, although 55% of the surviving patients in the TSR group reported being pain free compared to none in the HA group. A total of 4 revisions were performed in the HA group, and 2 in the TSR group.