Host and microbial characteristics associated with recurrent prosthetic joint infections

Abstract

AbstractApproximately 20% of patients after resection arthroplasty and antibiotic spacer placement for prosthetic joint infection develop repeat infections, requiring an additional antibiotic spacer before definitive reimplantation. The host and bacterial characteristics associated with the development of recurrent infection is poorly understood. A case‐control study was conducted for 106 patients with intention to treat by two‐stage revision arthroplasty for prosthetic joint infection at a single institution between 2009 and 2020. Infection was defined according to the 2018 Musculoskeletal Infection Society criteria. Thirty‐nine cases (“recurrent‐periprosthetic joint infection [PJI]”) received at least two antibiotic spacers before clinical resolution of their infection, and 67 controls (“single‐PJI”) received a single antibiotic cement spacer before infection‐free prosthesis reimplantation. Patient demographics, McPherson host grade, and culture results including antibiotic susceptibilities were compared. Fifty‐two (78%) single‐PJI and 32 (82%) recurrent‐PJI patients had positive intraoperative cultures at the time of their initial spacer procedure. The odds of polymicrobial infections were 11‐fold higher among recurrent‐PJI patients, and the odds of significant systemic compromise (McPherson host‐grade C) were more than double. Recurrent‐PJI patients were significantly more likely to harbor Staphylococcus aureus. We found no differences between cases and controls in pathogen resistance to the six most tested antibiotics. Among recurrent‐PJI patients, erythromycin‐resistant infections were more prevalent at the final than initial spacer, despite no erythromycin exposure. Our findings suggest that McPherson host grade, polymicrobial infection, and S. aureus infection are key indicators of secondary or persistent joint infection following resection arthroplasty and antibiotic spacer placement, while bacterial resistance does not predict infection‐related arthroplasty failure.