Wednesday, October 19, 2016

Failed total shoulder: what to do when the glenoid component is loose?


Revisions for aseptic glenoid component loosening after anatomic shoulder arthroplasty

These authors remind us that glenoid component loosening is a common indication for revision shoulder arthroplasty.

They report on 34 revision shoulder arthroplasties performed for aseptic glenoid loosening over two decades between January 1, 1985, and December 31, 2005.

Three patients were lost to follow-up. Treatment included component reimplantation in 20 shoulders (group I) or component removal with bone grafting in 11 shoulders (group II). 

Nine cases demonstrated instability with or without rotator cuff tearing prior to revision. 

The mean follow-up period was 8.3 years.

Pain relief occurred in 26 of 31 shoulders, with no difference between groups (P > .99). Active elevation and external rotation improved in both groups (P = .8). The rate of survival free of reoperation at 10 years was 78.9% in group I (red line) and 83.9% in group II (blue line) (P = .5).  



Five (25%) shoulders with glenoid remimplantation were recognized as having radiographically loose glenoids; two had reoperations. 

Nine shoulders without glenoid reimplantation had medial glenoid erosion, with two requiring reoperation for pain. 

There was a trend toward reoperation in those with preoperative instability (5 of 8 re-revisions).

The case example below shows the revision of a loose glenoid without reimplantation.

 Bone grafting of the defect was used.
 At 15 years after revision, the bone graft had resorbed, but the patient was satisfied with her outcome and had a stable articulation without risk of repeat failure of a prosthetic glenoid.




Comment: The landscape of loose glenoid components is changing. In the 'old days', many cases were truly 'aseptic'. But recent evidence demonstrates that as many as half of these cases will have positive cultures, often for Propionibacterium - which requires special culture protocols and culture observation for at least 2 weeks. The problem is that preoperative data and intraoperative findings (including gram stains and frozen sections) are not sensitive to the presence of this organism. Removing or revising a loose glenoid component and leaving the humeral stem in place in a shoulder that subsequently is found to be culture positive leaves a stem that most likely will have a Propionibacterium containing biofilm on it.

Here are two representative cases of glenoid loosening that were associated with no preoperative or intraoperative evidence of infection (negative blood work, gram stains, frozen sections). In both cases, multiple cultures were sent for Propionibacterium and in both cases >60% were strongly positive for this organism.

Case 1

Case 2





The point is that we cannot be sure whether a shoulder with a failed glenoid component is "aseptic" or not until the results of proper cultures are finalized. Intraoperative decision making about humeral component exchange and immediate post operative decisions regarding antibiotic therapy need to be made in the absence of the final culture results.

In the current era, we are inclined to treat cases of glenoid component loosening as infected until proven otherwise, with glenoid component removal without reimplantation or bone grafting, with humeral component exchange and with antibiotic coverage until the intraoperative cultures are found to have no growth.