Does retained cement or hardware during 2-stage revision shoulder arthroplasty for infection increase the risk of recurrent infection?

ABSTRACT

BACKGROUND: When treating chronic prosthetic joint infection after shoulder arthroplasty, removal of the implants and cement is typically pursued because they represent a potential nidus for infection. However, complete removal can increase morbidity and compromise bone stock that is important for achieving stable revision implants. The purpose of this study is to compare the rates of repeat infection after 2-stage revision for prosthetic joint infection in patients who have retained cement or hardware compared to those who had complete removal. MATERIALS AND METHODS: We retrospectively analyzed all two-stage revision total shoulder arthroplasties performed for infection at 2 institutions between 2011 and 2020 with minimum 2-year follow-up from completion of the two-stage revision. Patients were included if they met the International Consensus Meeting criteria for probable or definite infection. Postoperative radiographs after the first-stage of the revision consisting of prosthesis and cement removal and placement of an antibiotic spacer were reviewed to evaluate for retained cement or hardware. Repeat infection was defined as either ≥2 positive cultures at the time of second-stage revision with the same organism cultured during the first-stage revision or repeat surgery for infection after the two-stage revision in patients that again met the International Consensus Meeting criteria for probable or definite infection. The rate of repeat infection among patients with retained cement or hardware was compared to the rate of infection among patients without retained cement or hardware. RESULTS: Thirty-seven patients met inclusion criteria and were included in the analysis. Six (16%) patients had retained cement and 1 patient (3%) had 2 retained broken glenoid baseplate screws after first-stage revision. Of the 10 cases of recurrent infection, 1 case (10%) involved retained cement/hardware. Age at revision (60.9 ± 10.6 vs. 65.0 ± 9.6, P = .264), body mass index (33.4 ± 7.2 vs. 29.7 ± 7.3, P = .184), Charlson Comorbidity Index (2 (0-8) vs. 3 (0-6), P = .289), male sex (7 vs. 16, P = .420), and presence of diabetes (1 vs. 3, P = .709) were not associated with repeat infection. Retained cement or hardware was also not associated with a repeat risk of infection (1 vs. 6, odds ratio = 0.389, P = .374). DISCUSSION: We did not find an increased risk of repeat infection in patients with retained cement or hardware compared to those without. Therefore, we believe that surgeons should consider leaving cement or hardware that is difficult to remove and may lead to increased morbidity and future complications.