RESUMO
AIMS: Shoulder replacement is increasingly performed for end-stage arthritis. Information on the long-term survival and patient outcomes is very limited. This study aimed to quantify the survival and clinical outcome at a minimum of 20-yr follow-up. METHODS: A single-center, single-surgeon, consecutive cohort study was performed. Forty-four shoulder replacements in 40 patients (age at surgery 68.5 years, 82.5% female, preoperative visual analog scale [VAS] pain score 5.1/10, standard deviation [SD] 2.7) implanted between 1996 and 2000 were assessed. All-cause construct survival, radiographic glenoid and humeral stem loosening, radiographic humeral head migration, and patient-reported outcome measures were assessed. RESULTS: Survival with all-cause revision as an endpoint was 84.1% (95% confidence interval [CI] 60.7, 94.1) at 20 years, glenoid loosening was seen in all patients who survived to the 20-year follow-up. Survival of rotator cuff integrity was 16.8% (95% CI 3.5, 38.5) at 20 years. VAS pain scores demonstrated improvement at 10 years (mean change -4/10) but not at 20 years (effect size -0.15, mean change 0.4/10, SD 2.7). At 20 years, 72% of patients had died with the prosthesis in situ. CONCLUSION: Older patients undergoing total shoulder arthroplasty are unlikely to require revision in their lifetime. However, beyond 10 years, a large proportion of implants demonstrate glenoid loosening, humeral head migration, and declining patient outcomes. This information will be of use to patients and clinicians when discussing the potential outcomes of surgery.