There is nothing going to ruin your day like a complication after shoulder arthroplasty, either hemiarthroplasty (HA) anatomical (TSA) or reverse arthroplasty (RTSA). While complications are fortunately uncommon with anatomical shoulder arthroplasty (approximately 8% but as high as 40%), the complication rate for RTSA has been reported as high as 70%. Most complications are multifactorial and cannot all be blamed on the patient. Basically you do not want to operate upon a young patient who is an insulin dependent diabetic, has an ASA of 3 or 4, who smokes, has HIV disease and has a BMI over 40.

The most common predictors of failure are determined by the indication for surgery, the type of implant used and the skill of the surgeon. The major risk factors for HA are eccentric glenoids, young age and rotator cuff failure. The biggest risk factors for TSA are metal backed glenoid components, younger age of the patient, rotator cuff pathology and insufficient glenoid bone. The major risk factors for RTSA are the type of component used (Grammont type versus lateral center of rotation designs) and the indication for surgery. Infection risks include previous infected arthroplasty, previous joint infection, immunosuppression (e.g. steroid use, insulin dependent diabetes, HIV disease), ASA 3 or 4 and higher BMI (i.e. over 40). Not all of these risk factors can be prevented in patients but informed consent is critical to the patient understanding the potential outcomes of their surgery.

There are unfortunately many reasons a TSA can be painful after implantation, but the most common reason is sepsis. Making the diagnosis can be a major challenge, but the biggest challenge is to think of that as a diagnosis! The most important steps are to first obtain plain radiographs one week after surgery so that you can compare subsequent radiographs to the immediate post-operative films. Progressive radiolucent lines in the glenoid or especially around the humerus are important hints. A loose humeral component is infected until proven otherwise. Next blood work to include CRP and ESR are critical. Other markers of infection have not been used on a widespread basis. If there is concern that there might be rotator cuff pathology and not sepsis, then we obtain an arthrogram CT scan at the time of aspiration. A cell count is helpful but often there is a dry tap. It is important to create a “p. acnes protocol” at your hospital to take cultures out 15 days. If still not sure and revision is necessary then we aspirate the joint at the time of surgery for cell count. The WBC cell count at the time of surgery keeps changing but over 3000 is considered diagnostic. Multiple specimens are sent for frozen section, culture and permanent section. The more WBC per high powered field the more likely there is an infection. Gram stains are worthless and we do not rely on them. We have no experience with implant sonification or use of IL-6.

The current standard of care in the United States for the treatment of an infected arthroplasty, whether it be a TKA, THA or TSA remains two stage revision. We performed a systematic review of the English literature to answer the question of whether one-stage or two-stage revision of infected shoulder arthroplasty was better in terms of re-infection rate, patient satisfaction and patient function. There were no prospective, randomised studies comparing these two approaches. We found 19 studies with a minimum one year follow up which described 99 patients with two-stage revision, 38 with one-stage revision and 36 with spacers. There were no significant differences between the groups for recurrent infection rate or complication rate. The spacer only group had higher satisfaction than the one-stage group, but the satisfaction rate between the one- and two-stage groups was not statistically different.

Unfortunately, the numbers in our study were small and based upon level IV evidence. It is interesting that currently the data do not support two-stage over one-stage revision of an infected TSA. So why do we continue to use two-stage revision? I believe it is the fear of litigation to a certain degree but also historical bias toward two-stage revisions. The only time I would do a one-stage revision is if (1) it is a low virulence organism, (2) the prosthesis can be easily removed and (3) the patient was too sick to tolerate multiple operations.