Résumé
Background: Cutibacterium acnes prosthetic joint infection affects mainly, but not exclusively, the shoulder. There is limited or no data in the literature comparing clinical setting and failure rates between shoulder implants (SI) and lower limb implants (LLI).Method: comparative, multicentric, and retrospective study including monomicrobial Cutibacterium acnes prosthetic joint infections of the shoulder, knee, and hip occurring between 2012 and 2022 at the Montpellier and Nîmes University Hospitals. We compared clinical data and failure rates between SI infection and ILM. We also conducted an analysis of predictive factors for failure.Results: we included 56 patients. Pain was significantly more frequent and presentation more frequently acute in LLI. The most frequently antibiotic prescribed was clindamycin (>7 days in 87.5% of patients). Amoxicillin was used in 16.5% of patients. Rifampicin was associated in 28.6%. Failure rate was high, particularly for LLI (10/27, 37.0%) compared with 5/24 (20.8%) for SI (P-value=0.235). The reason for failure appeared to be different according to infected joint, with significantly more relapses with the same germ in SI (16.7% vs 0.0% P=0.043), and more new infections for LLI (29.6% vs 4.2% P=0.026).The only factor significantly associated with failure was acute infection. There was insufficient data to conclude on antibiotic therapy, but failures tended to be less frequent in patients who received amoxicillin (12.5%).Conclusion: the high frequency of new infections in failures in LLI infections suggests that the isolation of only C. acnes in LLI infection should be considered with caution. Our data are insufficient to suggest ways of optimising antibiotic strategy to improve the poor outcome of C. acnes PJI.