Abstract
Aim
Surgical and antimicrobial treatment of periprosthetic joint infections (PJI) depends largely on the causative pathogen. We assessed the pathogen detection rates and the concordance of preoperative synovial fluid culture and culture of intraoperative samples harvested during revision surgery in patients with PJI.
Method
Culture-positive PJI cases treated at our institution from 02/2011 to 07/2018, for which culture results from preoperative (synovial fluid) and intraoperative samples (periprosthetic tissue, synovial or sonication fluid) were available, were retrospectively assessed. For organisms belonging to the resident skin flora (coagulase-negative staphylococci, cutibacteria and corynebacteria) significant growth was considered, if the identical pathogen grew in ≥2 samples or >50 cfu/ml sonication fluid. For other pathogens (S. aureus, streptococci, enterococci, fungi and gram-negative rods) or patients under antimicrobials, any growth was considered positive. We determined the pathogen detection rate in preoperative and intraoperative cultures and compared it in different subgroups using Fisher's exact test. Furthermore, we assessed the concordance of preoperative and intraoperative cultures.
Results
We included 167 culture-positive PJI cases (76 hip and 91 knee joints). Coagulase-negative staphylococci (n=55, 33%), Staphylococcus aureus (n=34, 20%) and streptococci (n=22, 13%) were the most common pathogens. In 17 cases (10%) polymicrobial infection was found. The pathogen(s) grew in synovial fluid in 105 cases and in intraoperative samples in 146 cases (63% vs. 87%, p<0.001). 49 patients received antibiotics before aspiration and/or surgery. No differences were observed comparing hip and knee prostheses, primary and revision prostheses or patients receiving or not receiving antibiotics before sampling. Congruent results of preoperative and intraoperative cultures were found in 85 cases (concordance 51%). In 14 cases (8%), the pathogen was detected preoperatively only, in 59 cases (35%) the pathogen was found intraoperatively only; in 3 cases an additional pathogen was found preoperatively, in 6 cases an additional organism was found intraoperatively. Pathogen detection was significantly better in intraoperative compared to preoperative cultures in low-virulent pathogens (87% vs 36%, p<0.001), polymicrobial infections (88% vs. 47%, p<0.001) and delayed/late PJI (>3months; 92% vs 64%, p<0.001). There was no difference regarding detection rate of high-virulent pathogens (88% vs 83%) and in early postoperative PJI (<3 months, 91% vs. 73%).
Conclusions
As concordance of preoperative and intraoperative microbiological results was 51%, surgical and antimicrobial treatment should not be selected based on preoperative synovial fluid cultures only. An additional pathogen was found intraoperatively in 39%.
