We report our experience of revision shoulder arthroplasty at Wrightington Hospital. Thirty-Nine patients had undergone revision surgery and followed up for a minimum of two years. Patients were scored using the Constant score and the ASES score pre-operatively and post operatively. All patients had X-ray evaluation for loosening and migration. Of the thirty-nine patients, 16 were failed humeral head replacement (HHR) and 16 were failed total shoulder replacement (TSR). All but two of the HHR were revised for glenoid erosion to a TSR, there was an equal proportion of patients with rheumatoid arthritis and osteoarthritis. Of the 16 patients undergoing revision surgery for failed TSR 6 were rheumatoid, 4 had osteoarthritis and 5 had posttraumatic arthritis. The main reasons for revision include glenoid loosening (7) instability (4) and peri-prosthetic fracture (2). The average constant scores post operatively for HHR and TSR were 35.5 (sd+/− 21.1) and 29.1 (sd+/− 12.1) respectively. The average ASES scores for HHR and TSR were 60.5(sd +/ 27.8) and 50.1(sd +/− 22.0) respectively. There was no statistical difference between the two groups in respect to the constant scores (p value 0.18) or ASES scores (p value 0.16). Overall, the pain relief was good post operatively following both HHR and TSR. The mean visual analogue score for pain following HHR was 3.2 and following TSR 3.5. Range of movement, function and strength was poor following both HHR and TSR.

HHR fail in a predictable way and can be revised with conversion to a TSR. TSR fail in a variety of ways and there revision surgery is demanding and complex. Both types of revision offer good pain relief but poor function.