Abstract
Aim
Cutibacterium acnes, a skin commensal, is responsible for 5–10% of prosthetic joint infections (PJI). All current microbiological definitions of PJI require two or more identical commensal isolates to be recovered from the same procedure to diagnose PJI and rule out contamination. Unlike coagulase negative staphylococci, C.acnes shows a highly stereotypical susceptibility profile making impossible to phenotypically assess the clonal relationship of isolates. In order to determine the clonal relationship of multiple C.acnes isolates recovered from arthroplasty revisions, we analyzed by multi-locus sequence typing (MLST) C.acnes isolates grown from orthopedic device-related infections (ODRI) in a reference center for bone and joint infection.
Methods
Laboratory records from January 2009 to January 2014 were searched for monomicrobial C.acnes ODRI with growth of C. acnes in at least 2 intraoperative and/or preoperative samples. Clinical, biological and demographic information was collected from hospital charts. All corresponding isolates biobanked in cryovials (−80°C) were subcultured on anaerobic blood agar, and identification confirmed by MALDI-TOF-MS. C.acnes isolates were typed using the MLST scheme described by Lomholt et al. Plasmatic pre-operative C-reactive protein (CRP) levels were determined using DimensionEXL (Siemens). A threshold of 10 mg/L was used to determine serologically positive ODRIs from negatives.
Results
Over a 5-year period, 37 cases of monomicrobial C.acnes ODRI were diagnosed in our center. Among these 37 cases, 113/153 C.acnes isolates were cryopreserved. 110/113, corresponding to 36/37 cases, were typed by MLST: 14/36 (39%) ODRI cases were found to feature isolates belonging to two or more different STs and were qualified to be heteroclonal whereas 22/36 (61%) of ODRI cases were found to feature isolates belonging to the same ST and were qualified to be homoclonal. Homoclonal infections were significantly more likely to have elevated CRP levels compared to heteroclonal cases (p=0.0011, Fisher test). Patients with only two positive intraoperative samples had significantly lower CRP values than patients with three or more positive intraoperative samples (12,7mg/L vs 67mg/L; p=0,01, homoscedastic two-tailed Student's t test).
Conclusions
This study suggests that what is classified microbiologically as C.acnes ODRIs comprises: i) true homoclonal infections eliciting an inflammatory response, ii) heteroclonal infections lacking inflammatory response where C.acnes could be an innocent bystander and iii) false positives where no strain achieves true microbiological significance. Our study shows that a stricter threshold of 3 intraoperative positive samples could be more adequate than 2. These results reinforces the need for a more specific definition of C.acnes ODRI.
