Factors influencing recurrence rates and radiological outcomes after two-stage exchange arthroplasty for prosthetic joint infection utilizing a premade antibiotic loaded cement hemiarthroplasty.

Background: The purpose of this study was to describe the rate of infection recurrence in a patient cohort undergoing placement of a prefabricated antibiotic spacer in the setting of two-stage management of shoulder prosthetic joint infection (PJI). Methods: All patients who underwent placement of a prefabricated antibiotic cement spacer for infection by three surgeons between 2019 and 2024 with a minimum of 3-month follow-up after spacer placement were retrospectively reviewed. Subsequent infection was defined as development of a sinus tract, purulent drainage, or subsequent revision surgery-planned or unplanned- associated with ≥2 tissue specimens with growth of the same bacterial species taken during a biopsy prior to the second stage if performed, at the second stage if performed, or after the second stage during revision of the second stage. Patients were characterized by the International Consensus Meeting 2018 criteria, and comparisons were performed between criteria and reinfection. Radiographic review was performed analyzing spacer loosening, fracture, and glenoid wear. Reinfection and nonreinfection groups were compared. Results: A total of 55 shoulders (54 patients) met the inclusion criteria. The mean age and follow-up were 61.1 ± 11.1 years and 14.7 ± 10.8 months, respectively. Overall, 38.2% (21/55) of patients experienced reinfection. Five patients had clinically reinfected spacers, 7 had positive cultures on a planned prerevision biopsy, 7 had clinical reinfection after the second stage, and 2 had positive cultures at the time of second stage. Patients with a definite infection per International Consensus Meeting shoulder criteria demonstrated a 50% reinfection rate compared to 33% for probable, 23% for possible, and 33% for unlikely (P = .349). Patients with reinfections were more likely to be younger patients (55.4 ± 11.7 vs. 64.6 ± 9.2; P = .002) and have longer follow-up (19.5 ± 12.1 vs. 11.7 ± 8.9 months; P = .015). The most common culture results were culture negative (26/55; 47.3%) and Cutibacterium acnes (18/55; 32.7%). Only 14.3% (3/21) of patients that developed reinfection were found to have the same microbe on subsequent cultures. Radiographic factors including spacer fracture and glenoid wear pattern were not associated with recurrent infection (P > .05). Conclusion: Patients undergoing two-stage revision for PJI with a prefabricated antibiotic cement spacer demonstrate a high rate of recurrence (38.2%), using our conservative definition of recurrence and often become reinfected with a new bacterial species (85.7%). Patient factors are also important to consider given that recurrently infected patients were younger and had longer follow-up. A standardized definition of recurrent PJI is mandatory to accurately compare studies utilizing various treatment options.