As shoulder arthroplasty becomes more common, the number of failed arthroplasties requiring revision is expected to increase. When revision arthroplasty is not feasible, resection arthroplasty has been used in an attempt to restore function and relieve pain. Although outcomes data for resection arthroplasty exist, studies comparing the outcomes after the removal of different primary shoulder arthroplasties have been limited.This was a retrospective multicenter review of 26 patients who underwent resection arthroplasty for failure of a primary arthroplasty at a mean follow-up of 41.8 months (range, 12-130 months). Resection arthroplasty was performed for 6 failed total shoulder arthroplasties (TSAs), 7 failed hemiarthroplasties, and 13 failed reverse TSAs.Patients who underwent resection arthroplasty demonstrated significant improvement in visual analog scale pain score (6 ± 4 preoperatively to 3 ± 2 postoperatively). Mean active forward flexion and mean active external rotation decreased, but this difference was not significant. Subgroup analysis revealed that postoperative mean active forward flexion was significantly greater in patients undergoing resection arthroplasty after failed TSA than after reverse TSA (P = .01).Resection arthroplasty is effective in relieving pain, but patients have poor postoperative function. Patients with resection arthroplasty for failed reverse shoulder arthroplasty have worse function than those with failed hemiarthroplasty or TSA. Surgeons should be aware of this when assessing postoperative function. There is no difference in functional outcome between hemiarthroplasty and TSA.