We present a series of 18 consecutive cases of primary reverse total shoulder arthroplasty for irreparable proximal humerus fractures in patients over 70. Failure of tuberosity union and poor cuff function leads to unsatisfactory results in over half the patients with hemiarthroplasty. Reverse total shoulder arthroplasty does not depend upon a functional rotator cuff and requires little formal rehabilitation.

Patients over 70 with irreparable proximal humerus fractures treated with a reverse total shoulder arthroplasty were included in this study. Only primary arthroplasties were included. Reverse arthroplasties for failed hemiarthroplasties were excluded.

All arthroplasties were performed using either a deltoid split direct lateral (superior) approach or the antero-lateral MacKenzie approach. The SMR reverse total shoulder prothesis was implanted in all cases using a press-fit glenoid base plate and glenosphere, and press-fit or cemented humerus stem. Tuberosity repair was attempted in 10 cases. The supraspinatus was excised from the greater tuberosity.

Patients were allowed self-mobilisation after two weeks in a sling. Patients were recruited and followed up per ethics approved protocol.

Outcome measures used were range of motion, dislocation and revision rates radiological signs of loosening and glenoid notching, DASH and Constant scores. Results were compared to another series of cases of reverse shoulder arthroplasty for sequelae of trauma and failed hemiarthroplasties, as well as a series of primary hemiarthroplasties.

At an average follow-up of 30 months (minimum 12 months) all patients were satisfied with their results. Average forward elevation was 132 deg. and abduction 108 deg. There was not deterioration of movement at 12 or 24 months. No patient had ongoing pain. The average constant score was 62.

There was no evidence of humeral stem loosening apart from one case of early subsidence in a press fit stem. Eleven cases showed glenoid notching, four Nerot grade 1, six Nerot grade 2 and one Nerot grade 3. All notching had stabilised after 12 months. There were no cases of dislocation. No case needed revision, or awaits revision. All cases were pain-free at last review.

Overall results for this group of primary reverse arthroplasties for fractures was much better than for reverse arthroplasties for sequelae of trauma. The results were also better than for primary hemiarthroplasties. Irreparable three and four part fractures of the proximal humerus pose management challenges in the elderly. The reverse total shoulder arthroplasty is very attractive option for elderly patients with irreparable proximal humerus fractures. They require little rehabilitation and can give reproducibly good functional results, which do not deteriorate with time.