Low sensitivity of periprosthetic tissue PCR for prosthetic knee infection diagnosis

We previously described a rapid PCR panel targeting bacteria associated with prosthetic joint infection (PJI) for use on biofilms dislodged from explanted hip and knee arthroplasties. Herein, we tested the PCR panel on periprosthetic tissues from 95 subjects undergoing resection knee arthroplasty, all of whom had had tissue cultures performed. A subset also had synovial fluid culture (n = 89) and/or device vortexing and sonication with culture (n = 58), and a subset of these PCR (n = 36) of resulting sonicate fluid, performed. Of the 64 subjects with PJI, 10 and 44 had positive tissue PCR and cultures, respectively, one of whom had a positive tissue PCR with negative tissue cultures. The overall sensitivity of tissue PCR (16%) was much lower than that of tissue (69%), synovial fluid (72%) and sonicate fluid (77%) culture as well as sonicate fluid PCR (78%) (P < 0.00001, P < 0.00001, P < 0.00001, and P = 0.0003, respectively). Periprosthetic tissue PCR, as performed herein, has poor sensitivity for PJI diagnosis.