Aims: In this study we assessed the results of the hemi-arthroplasty for shoulder fractures in patients with a follow-up 2 to 7 years. Methods: Twenty six patients (20 women and 6 men, mean age 64.7±8.2 years [range 41 to 78 years]), with a fracture of the upper humerus requiring hemiarthroplasty, were followed-up 2–7 years after surgery. Ten Coþeld, nine Global and seven Aequalis prosthesis were implanted, all cemented. All the procedures were performed 0–17 days after the injury (mean 5.5±4.6 days). The clinical outcome was assessed using the Constant-Murley scale. Results: The mean score, at their last follow up, was 70.4±16.4% (39–96%). The mean ßexion of the arm was 150û (30û–175û), the mean abduction was 145û (30û–170û), the mean external rotation was 30û (10û–45û) and the internal rotation corresponded with a position of the dorsum of the hand at the L3 vertebrae. The patients in our series achieved the optimum clinical result during the þrst 6 months after the operation. No statistical signiþcant improvement occurred after this period. Six months after the injury 20 out of the 26 patients (76.9%) had the same activity level as they had prior to the fracture. At their last follow up eighteen patients (69.2%) had no any pain and 7 (27%) patients had some mild pain at the end of their daily activities and 1 patient (3.8%) had pain even with mild activities. Conclusions: Shoulder hemiarthroplasty seems to be a worthwhile procedure in modern orthopaedic surgeonsñ armamentarium, giving predictable results presuming careful selection of the patients, restoration of the individual anatomy of the shoulder and aggressive rehabilitation program during the þrst 6 months after surgery.

Aims: The objective of this study was to assess whether anatomical placement of the prosthesis, in shoulder hemi-arthroplasty for fracture of the humeral head, is important to the clinical outcome. Methods: Sixteen patients, with a fracture of the upper humerus treated with hemi-arthroplasty, were followed-up 45.7±15.1 (20–72) months after surgery. The results were assessed using the Constant-Murley scale. The mean score was 75.8±15.7% (54–96%). At the time of their last follow up they underwent CT of the fractured and sound humerus, in order to be measured differences in humeral length and retroversion using special software. Correlation between these differences and the clinical outcome, as it was measured with the Constant-Murley scale, was performed. Results: The mean difference in retroversion was 8.7 degrees and the mean difference in length was 0.65cm, between fractured and sound humerus in our patients. We have achieved a very good þnal outcome (Constant score more than 71%) in patients with difference in retroversion less than 10 degrees and difference in length less then 14mm, between fractured and sound humerus. Conclusions: Restoration of the humeral length and retroversion is very important in shoulder hemiarthroplasty for fracture of the humeral head. Only small differences from the optimum length are well tolerated while only big differences from the optimum retroversion are likely to affect signiþcantly the clinical outcome. We attribute the very good clinical outcome in our series to the quality of the anatomical reconstruction that was performed despite the fact that our sample is small and we cannot have powerful statistics.