The is more limited surgical outcome data on shoulder PJI compared to the hip and knee and even less when specifically examining the use of RSA in the setting of infection. Small sample size, limited follow-up, and varying treatment protocols can limit the conclusions drawn from these studies. However, with the growing use of RSA in the last decade, there will be an increase in the reported outcomes for this patient population.

While less common than a two-stage exchange, outcomes for one-stage revision have been reported in the literature. Beekman et al. [15] report on 11 cases of one-stage revision for infected reverse prostheses. All patients were revised to a cemented RSA with antibiotic-impregnated cement. The isolated organisms were P. acnes (7 shoulders), coagulase-negative Staphylococcus (5), methicillin-resistant Staphylococcus aureus (1), and Escherichia coli (1), including two multibacterial infections. A minimum of 3 days of IV antibiotic therapy was given to all patients, and the minimum overall course of antibiotic treatment (combined IV and oral) was 3 months. All but one patient was considered free of infection after one-stage revision at a mean follow-up of 24 months, and without antibiotic treatment for a minimum of six months. The one persistent infection was ultimately cleared after a second revision for placement of an antibiotic prosthetic spacer. Overall, in this series, the mean Constant score was 55 at final follow-up. The authors concluded that one-stage revision arthroplasty can be reliable for eradicating infection in RSA, with average functional outcomes and a reduction in cost and duration of treatment compared to a two-stage procedure.

More recently, Klatte et al. [16] retrospectively examined 35 patients who underwent a single-stage exchange procedure for an infected shoulder arthroplasty. Nine of these patients received a RSA, with the others receiving a hemiarthroplasty (18), or a hemiarthroplasty with a bipolar head (8). Outcomes were broken down by final prosthesis. All procedures were a planned single-stage exchange with the preoperative identification of a specific organism and its sensitivities by joint aspiration and culture. Cement fixation was impregnated with the appropriate antibiotic, and specific IV antibiotics were started during surgery. Appropriate IV antibiotic therapy was administered for a mean of 10.6 days postoperatively (range 5–29 days). The most commonly isolated organism was coagulase-negative Staphylococcus (11) and P. acnes (12). In this study, 33/35 patients (94 %) did not develop a further infection at a mean follow-up of 4.75 years. The best outcome as judged by the Constant–Murley score was for the RSA group (61, range 7–90), although this difference was not significant when compared to the hemiarthroplasty group (43.3 range 14–90), and the hemiarthroplasty with bipolar group (56 range 40–88). The authors conclude that a single-stage exchange gives a high rate of success for eradication of the infection, although a careful pre-, intra-, and postoperative workup and treatment plan is required.

A number of studies have looked at outcomes with use of RSA following two-stage exchange for infection. Hattrup and Renfree [17] reported on a series of 25 patients undergoing two-stage exchange for the treatment of glenohumeral sepsis, for which 20 patients had an infected shoulder arthroplasty, and 8 had placement of a RSA. The most common isolated bacteria were coagulase-negative Staphylococcus (8 shoulders), P. acnes (7), and methicillin-sensitive Staphylococcus (3). Organism-specific IV antibiotic therapy was administered for all patients for 6 weeks between explant and reimplantation. Overall, pain relief and improved function were noted at final follow-up, with mean active forward flexion increased from 47° to 100.9° (p < 0.001), mean abduction increased from 38.8° to 93.6° (p < 0.001), and mean external rotation increased from 6.8° to 32.6° (p = 0.0012). A nonsignificant increase in overhead function was seen in patients with a RSA (124.4° active flexion versus 87.9° in anatomic arthroplasty), but pain relief (VAS) and other functional scores (SST—8.6 vs. 7.7, ASES—69.9 vs. 67.4) were not different between reverse TSA and anatomic arthroplasty. There were 3 recurrent infections, all from P. acnes. Two of the RSA patients had postoperative dislocations that required revision to retentive polyethylene liners.

Coffey et al. [18] reported on their experience with two-stage exchange for infected shoulder arthroplasty and native septic arthritis for 16 shoulders (11 shoulder prostheses; 6 hemis, 3 reverse TSAs, 2 standard TSA) using a commercially produced antibiotic-impregnated spacer. Nine of the shoulders were reimplanted to a reverse TSA. All patients underwent placement of a commercially manufactured gentamycin-impregnated spacer and received culture-specific IV antibiotic therapy postoperatively (mean IV treatment 5.6 weeks). Reimplantation occurred at a mean of 11.2 weeks (range, 6–30) after implant removal and spacer placement based on normalizing serum IL-6 levels. Mean follow-up was 20.5 months (range, 12–30). Outcomes were measured independent of implant selection at final revision. For the entire 16 patient cohort, pain was improved, with mean active forward flexion increased from 65° before spacer placement to 110° at final follow-up and mean active external rotation increased from −5° to 20°. At final follow-up, mean functional scores for the UCLA, SST, ASES, and Constant score were 26, 6.6, 74, and 57, respectively. There were no recurrent infections.

Sabesan et al. [19] evaluated the outcomes of 17 patients who underwent a two-stage exchange in the treatment of infected shoulder arthroplasty, in which reimplantation was with a reverse TSA. Of these, 10 initially had a hemiarthroplasty, 4 had a TSA, and 3 had a RSA. Once again, Staphylococcus species (7 shoulders) and P. acnes (5) were the most commonly isolated organisms. Patients received organism-specific IV antibiotic therapy for a mean of 6.3 weeks (range, 4–54) postoperatively and had a median of 4.0 months (range, 1.8–61) between explant and reimplantation. Mean follow-up was 46.2 months (range, 22–80), and there was one recurrent infection from P. acnes that was ultimately cleared with a second two-stage exchange, and good functional outcomes were achieved. Mean Penn Shoulder score was significantly improved from preoperative levels at final follow-up (24.9–66.4), with mean forward flexion of 123° and mean external rotation of 26°. There were seven complications postoperatively, requiring 7 additional surgeries. This included 5 polyethylene exchanges or revision of glenosphere component for instability, 1 irrigation and debridement for postoperative hematoma, and 1 two-stage exchange for the recurrent infection. The authors concluded that two-stage reimplantation with a RSA has a low rate of recurrent infection and can improve pain and function even when major bone or soft tissue deficits, such as the rotator cuff, are present.

Jawa and colleagues examined 27 patients with an infection at the site of a shoulder arthroplasty who were managed with prosthesis of antibiotic-loaded acrylic cement (PROSTALAC) [20]. Sixteen patients underwent the second-stage procedure (11 declined), for which 10 of these had placement of a RSA. The mean duration of follow-up was 27.6 months (range 12–69 months). There were 5 (5/27) recurrent infections for the entire patient cohort and 2 (2/10) recurrent infections for patients who underwent placement of a RSA. Other complications for the entire cohort included 3 PROSTALAC fractures and 1 dislocation. At final follow-up, the average forward elevation was 77°. The authors noted that patients with RSA did not have improved functional outcomes compared to the other cohorts.

Cuff et al. [21] compared the results of one- and two-stage exchange in the treatment of deep shoulder infections for 22 patients, with revision to a reverse TSA for all patients with minimum 2-year follow-up. Of the 22 patients, 17 cases were treated for an infected shoulder hemiarthroplasty and 5 cases of infection following open rotator cuff repair. Ten shoulders underwent a one-stage exchange with reimplantation of a reverse TSA using antibiotic-impregnated cement, and twelve shoulders underwent a two-stage exchange with placement of a temporary antibiotic-impregnated spacer prior to reimplantation with a reverse TSA using antibiotic-impregnated cement. The mean time to reimplantation for the two-stage group was 16.2 weeks (range, 9–28). The final decision to proceed with one- or two-stage revision was based on clinical and intraoperative signs of infection including preoperative markers of infections, a subjective assessment of the quality of the debridement, and the accompanying pathology. Half (5/10) of the patients undergoing one-stage exchange had no clinical or intraoperative signs of infection, compared to only 1/12 undergoing the two-stage exchange. In this series, Staphylococcus species were the most commonly isolated organism (11 cases), while P. acnes was found in only 2. Patients received postoperative IV antibiotic therapy of 2- to 6-week duration. Mean follow-up was 43 months (range, 25–66). There were no recurrent infections for both treatment groups, and other outcome measures significantly improved from preoperative levels at final follow-up. There was improvement in mean abduction (36.1°–75.7°), forward flexion (43.1°–79.5°), and external rotation (10.2°–25.4°). Mean ASES score improved from 31.9 to 57.0, mean VAS pain score improved from 6.3 to 3.5, and the mean SST score improved from 1.3 to 4.0. Fourteen shoulders were rated as good or excellent, five as satisfactory, and three as unsatisfactory. No significant differences were seen between groups in regard to any of the functional outcomes. Eleven complications developed postoperatively in seven shoulders (32 %), requiring 5 additional surgeries. These included 3 patients with postoperative hematomas requiring irrigation and debridement, 1 patient with early dislocation treated with closed reduction, 1 patient with loosening of the humeral and glenoid component requiring revision to a long-stemmed reverse TSA, 1 fracture and dissociated polyethylene liner from a fall requiring revision surgery, and 1 other humeral fracture distal to the tip of the antibiotic spacer requiring ORIF with subsequent placement of a reverse TSA [21].

In a study by Ortmaier et al. [22], retrospective analysis of 20 patients who were treated for infection following RSA was performed. The majority of patients were treated with either a two-stage revision with RSA placement (12) or an irrigation and debridement with exchange of the polyethylene liner (7), while one patient underwent resection arthroplasty. The most frequent organisms were coagulase-negative Staphylococcus and P. acnes. The study reported a poor eradication rate for acute or subacute infection for patients who received irrigation and debridement with polyethylene liner exchange. Infection was not eradicated in 5/7 patients, with the majority undergoing a subsequent two-stage revision. In contrast, for the initial two-stage revision group, only 3 of 12 patients had a persistent infection. Postoperative instability without evidence of infection was present in an additional two patients who underwent the two-stage revision. In the 12 patients who underwent revision RSA, mean Constant score was 52.2 (range 28–67), mean UCLA score was 23.3 (range 18–32), mean SST was 6.4 (range 3–8), and mean VAS was 1.3 (range 0–4). The study reports improved outcomes for patients who underwent the initial two-stage revision for an infected RSA. However, only one-third of those patients who required two-stage revision for deep infection were pain-free at the time of follow-up.