The indications for surgical treatment of displaced proximal humerus fractures in older patients have gradually changed in the past 10 years: fewer hemiarthroplasty (HA) procedures but more reverse shoulder arthroplasty (RSA) procedures are being done. The aim of this study was to determine the clinical and radiological outcomes, the complications, reoperations and revisions of RSA and to compare them with those of HA.

A systematic review of the literature published between January 2006 and February 2016 was carried out to identify studies reporting or comparing results of HA and RSA for proximal humerus fractures in patients above 65 years of age. These studies were classified based on their quality and level of evidence. Of the 67 studies that were identified and reviewed, 22 were chosen for the analysis.

The functional outcome (Constant score) after RSA is significantly better and more reproducible that the one obtained after HA. RSA provides significantly better active range of motion in forward flexion and abduction than HA. Conversely, active internal and external rotation are worse after RSA than HA. Reattachment of the tuberosities around the RSA improves the rotation ability. Even if the tuberosities do not heal, the functional outcomes are satisfactory after RSA but not HA due to a major functional deficit. The tuberosity healing rate around the RSA does not decline with age, contrary to HA where age is a negative factor. The overall complication rate is higher after RSA than HA; however the reoperation rate is equal. The revision rate for implant change is higher after HA.

Compared to HA, RSA provides more reproducible function with better recovery of active forward flexion and abduction, even when the tuberosities are not reattached or do not heal properly. The poor rotation ability can be improved by reinserting the tuberosities. While the complication rate after RSA is higher than after HA, the reoperation rate is equal, and the revision rate is lower.

IV, Systematic review of literature.