LAUSR

BACKGROUND Our aim was to analyze the epidemiology, etiologies, and revision options for failed shoulder arthroplasty from 2 tertiary centers. METHODS From 1993 to 2013, 542 failed arthroplasties were revised in 540 patients (65% women): 224 hemiarthroplasties (HAs, 41%), 237 anatomic total shoulder arthroplasties (TSAs, 44%) and 81 reverse total arthroplasties (RSAs, 15%). Data about patients, pathology, and reintervention procedures, as well as intraoperative data, were analyzed from our 2 local registries that prospectively captured all the revision procedures. Patients had an average follow-up period of 8.7 years. RESULTS The revision rate was 12.7% for HAs, 6.7% for TSAs, and 3.9% for RSAs. HAs were revised earlier (33 ± 40 months) than RSAs (47 ± 150 months) and TSAs (69 ± 61 months). Glenoid failure was a major cause of reintervention: erosion in HAs (29%) or loosening in TSAs (37%) and RSAs (24%). Instability was another major cause of reintervention: 32% in RSAs, 20% in TSAs, and 13% in HAs. Humeral implant loosening led to revision in 10% of RSAs, 6% of HAs, and 6% of TSAs. Multiple reinterventions were required in 21% of patients, mainly for instability (26%) and/or infection (25%). The final implant was an RSA in 48%, especially when associated with cuff insufficiency, instability, and/or bone loss. Final reimplantation was possible in 90% of cases, with the remaining 10% treated with a resection or spacer. CONCLUSION Glenoid failure and instability are the most common causes of revision. Soft-tissue insufficiency and/or infection results in multiple revisions. Surgeons must recognize all complications so that they can be addressed at the first revision operation and avoid further reinterventions. RSA was the most common final revision implant.