BACKGROUND:
Avascular necrosis (AVN) of the humeral head frequently results in humeral head collapse and end-stage arthritic changes of the glenohumeral joint. Despite the recent proliferation of reverse total shoulder arthroplasty (RTSA), reports on the use of RTSA for AVN remain limited. The purpose of this study was to document the outcomes of shoulders indicated for RTSA in the setting of humeral head AVN and compare these with AVN shoulders indicated for the gold standard, anatomic total shoulder arthroplasty (aTSA).

METHODS:
A retrospective review of a multinational shoulder arthroplasty database was performed between August 2005 and August 2017. All shoulders with a preoperative diagnosis of AVN (aTSA in 52 and RTSA in 67) were reviewed. The shoulders in the RTSA cohort were matched (1:1) to shoulders with cuff tear arthropathy, whereas the shoulders in the aTSA cohort were matched (1:1) to shoulders with primary osteoarthritis. The mean follow-up period was 47 months (range, 24-130 months) for RTSA and 54 months (range, 24-124 months) for aTSA. Shoulders were evaluated for active range of motion (ROM) and patient-reported outcome measures (PROMs) prior to surgery and at latest follow-up. Patients treated with RTSA were compared with both the aTSA study cohort and the control group using the Student t test or χ2 test as indicated.

RESULTS:
RTSAs performed for AVN demonstrated significant improvements in all ROMs and PROMs. Patients undergoing aTSA for AVN were significantly younger than those undergoing RTSA (59 years vs. 73 years, P <  .001). At similar follow-up points, the RTSA cohort demonstrated significantly greater improvement in abduction (+51° vs. +32°, P = .03) whereas the aTSA cohort demonstrated significantly greater improvement in internal rotation. Postoperative University of California, Los Angeles scores (30 vs. 27, P = .014) and visual analog scale scores (1.4 vs. 2.4, P = .025) were better after RTSA; however, these differences between prosthesis types did not exceed the minimal clinically important difference. When compared with the control patients, the patients undergoing RTSA for AVN showed similar improvements in all ROMs and PROMs. Similarly, aTSA performed for AVN resulted in comparable improvements in pain, ROMs, and PROMs compared with aTSA performed for primary osteoarthritis.

CONCLUSION:
RTSA results in similar PROMs to aTSA in the treatment of AVN. Therefore, surgeons should continue to consider other patient factors such as glenoid bone loss and rotator cuff status when selecting implant polarity in patients with AVN.