© 2020 Elsevier Ltd Background: Shoulder replacement is an increasingly common treatment for end-stage degenerative shoulder conditions. Some shoulder replacements are unsuccessful and additional operations might be required. It is important for patients and clinicians to know how long shoulder replacements last and how effectively they reduce pain and improve function. This study aims to determine the longevity and long-term efficacy of shoulder replacements. Methods: In this systematic review and meta-analysis, we searched MEDLINE and Embase from their inception to Sept 24, 2019, for case series and registry data reporting 10-year or longer survival of total shoulder replacements, humeral hemiarthroplasties, and reverse total shoulder replacements of a specific brand of implant. Survival, implant, and patient-reported outcome measures data were extracted. The primary outcome was implant survival. We reviewed and analysed national joint replacement registries separately. We weighted each series by SE and calculated a pooled survival estimate at years 10, 15, and 20. For patient-reported outcome measures we pooled the standardised mean difference at 10 years. This study is registered with PROSPERO, CRD42019140221. Findings: 927 non-duplicate articles were identified by our search, nine articles (reporting ten series) were eligible for analysis of survival and six articles were eligible for analysis of patient reported outcomes. The ten series reported all-cause survival of 529 total shoulder replacements and 364 humeral hemiarthroplasties; no series for reverse total shoulder replacement met the inclusion criteria. The estimated 10-year survival for total shoulder replacement was 94·6% (95% CI 93·6–97·6) and humeral hemiarthroplasties was 90·4% (87·0–94·0). A single registry contributed 7651 total shoulder replacements, 1395 humeral hemiarthroplasties, and 7953 reverse total shoulder replacements. The pooled registry 10-year survival was 92·0% (95% CI 91·0–93·0) for total shoulder replacement, 85·5% (83·3–87·7) for humeral hemiarthroplasties, and 94·4% (93·4–95·7) for patients with osteoarthritis who had reverse total shoulder replacement and 93·6% (91·1–95·8) for patients with rotator cuff arthropathy who had reverse total shoulder replacement. Pooled 10-year patient-reported outcome measures showed a substantial improvement from baseline scores, with a standardised mean difference of 2·13 (95% CI 1·93–2·34). Interpretation: Our data show that approximately 90% of shoulder replacements last for longer than 10 years and patient-reported benefits are sustained. Our findings will be of use to surgeons and patients in the informed consent process and to health-care providers for resource planning. Funding: The National Institute for Health Research, the National Joint Registry for England, Wales, Northern Ireland, and Isle of Man, and the Royal College of Surgeons of England.
The lancet rheumatology
e539 - e548