Patient Satisfaction and Complications Following Revision of Failed Shoulder Replacement

Jon J.P. Warner, MD, Maggie Coats-Thomas, BS, & Megan Marshall, BA – November 2015


Failure of the rotator cuff with loosening of the socket required a revision to a reverse prosthesis.

  1. Infection: The incidence of infection after shoulder replacement is probably higher than reported in the literature and most surgeons don’t keep a record of their infection rate. The reported infection rate varies from 0-4%. Our annual infection rate is about 0.4% (Click here to see the 2014 Boston Shoulder Institute Quality Reports.)
  2. Tendon Tear: Tendons that tear are never normal and tearing of the rotator cuff tendons may occur due to poor quality tissue, technical issues of the way the shoulder replacement is placed, or with trauma such as a fall following surgery. The incidence of rotator cuff tendon-tearing parallels glenoid loosening and not surprisingly becomes more frequent over time. That said, the survival of the shoulder replacement without need for a revision surgery is about 92% after 10 years and the subsequent risk of failure requiring revision is about 1%/year.
  3. Nerve problems: Nerve problems are usually the result of stretching at the time of surgery and not direct injury by surgery. The reported incidence is low but this is probably because most nerve injuries are temporary and not noticed by the surgeon or the patient.
  4. Glenoid (socket) loosening: Loosening of the socket can be the result of technical issues at time of the original surgery and is often associated with rotator cuff tendon tearing. Moreover, it is always important to rule out an infection as a cause of loosening of the socket. Younger males have a higher rate of socket loosening than females, and this is probably due to a higher activity level after their shoulder replacement surgery.
  5. Instability: Instability after conventional shoulder replacement is actually relatively rare and when it occurs is usually due to failure of the subscapularis tendon (which is removed and then repaired during the replacement surgery); it may also be due to failure to properly position the glenoid socket and the humeral component.
  6. Stiffness: Some degree of stiffness is expected after conventional shoulder replacement and this is mostly dependent on how stiff the shoulder was before surgery. If a shoulder has been stiff for many years due to arthritis, it is not realistic to expect surgery can reestablish flexibility lost for many years in tendons and ligaments around the shoulder. Most patients are not troubled by this due to marked pain relief that they experience after surgery. Some, however, may be disappointed with their functional motion. Most patients are very satisfied with their postoperative motion.
  • 98% of patients report that they are happy with the outcome of their surgery when asked 1 year after their procedure.
  • Unfortunately, failures of shoulder replacement surgery do occur and most of our experience has been in revision of total shoulder replacements performed elsewhere. This module describes our experience with revision of conventional shoulder replacement mostly to Reverse Shoulder Replacement.


Unstable Shoulder Replacement. This surgery was revised with a soft-tissue repair.  It is very rare to solve such a problem without changing the prosthesis to a reverse prosthesis.


This patient presented with a failed rotator cuff and painful loose glenoid socket. He was found to have an infection when an arthroscopic biopsy was performed. We frequently perform an arthroscopy in order to determine if an infection is present prior to doing the definitive revision to a new shoulder replacement. About 32% of all patients who presented to Dr. Warner with a failed shoulder replacement and who underwent a biopsy in the past 4 years were found to have an infection.


This patient had an infected and painful shoulder replacement based on an arthroscopic biopsy. The revision surgery was removal of the infected shoulder replacement and placement of an antibiotic cement spacer called a PROSTALAC as well as treatment with intravenous antibiotics. This patient actually had no pain after surgery and has not required additional surgery for 8 years. Most patients will require a definitive revision to a reverse shoulder replacement.



This patient had a loose glenoid socket (arrows) with failure of the rotator cuff (the humeral head is high up on the socket), and revision surgery to a reverse shoulder replacement required a bone graft (from the right hip, iliac crest) in order to build up support for the new reverse socket (base plate).


  • From 2011-2014, Dr. Jon J.P. Warner did 428 shoulder replacements. Of these 428 cases, 172 of them (41%) were reverse shoulder replacements and 256 of them (59%) were total shoulder replacements.


  • Of the 428 shoulder replacements performed from 2011-2014, 62 (14.5%) were revision cases.


  • Of the 62 revision cases:
    • 18 were revisions to total shoulder replacements
    • 2 were revisions to PROSTALAC spacers, which were retained permanently
    • 42 were revisions to reverse shoulder replacements


  • Of the 62 patients, we were able to assess patient satisfaction in 50 cases
    • 86% of these patients were happy with the outcome of the surgery and/or would do it again
    • 14% of these patients were unhappy with the outcome of the surgery and/or felt worse



  • 11/62 (17%) patients undergoing revision surgery had a complication. Most of these did not affect the patients’ final outcome and satisfaction.
    • 2/11 complications resolved without additional treatment
    • 6/11 remained satisfied despite the complication
  • 8/62 (12.9%) patients still had pain
  • 4 patients had loosening of the revision shoulder replacement (2 humeral stems and 2 sockets)
  • The overall reoperation rate was 24%
  • The Surgical Complications were:
    • 1 Axillary vein injury that was repaired without a problem for the patient. The patient then underwent a second revision surgery to a reverse prosthesis and did well.
    • 1 nerve dysfunction which was a stretch injury which fully recovered
    • 2 (3.2%) infections after surgery (These patients had an infection with their original surgery before their revision surgery)
    • 1 dislocation which underwent a revision surgery and then the patient did well
  • The Non-Surgical Complications were:
    • 1 Pulmonary Embolism which fully recovered
    • 1 lower extremity numbness which fully recovered


  • The incidence of shoulder infection after conventional shoulder replacement averages 0.98% (based on a review of the literature from 2002-2010).


  • The rate of loosening of the glenoid socket after a total shoulder replacement = 1.2%/year….so the likelihood of having a good functioning shoulder after 10 years is 88% and after 20 years it is 76%.
  • The likelihood of needing a revision surgery is 0.8%/year so after 10 years 92% of patients will not have needed a revision surgery and after 20 years 84% will not have needed another surgery.



  • More experienced surgeons who do more surgery have better results based on less blood loss, fewer complications, and shorter hospital stays. They also save the hospital money.


  • Arthroscopic tissue biopsy had 100% sensitivity, specificity and predictive value compared to X-ray guided aspiration, which had only 17% sensitivity and 58% negative predictive value.