Abstract
Aim
Unexpected negative-cultures (UNC) are a common diagnostic problem in periprosthetic joint infection (PJI) of the hip and knee when using culture-based methods. A novel molecular approach (MC)1 based on the identification of the vast majority of bacterial species in a single assay using species-specific bacterial interspacing region length polymorphisms and phylum-specific 16S rDNA sequence polymorphisms has demonstrated clinical utility in PJI diagnostics (1). In addition, MC provides an estimate of the leukocyte concentration in the specimen analysed. The aim of this retrospective, blinded study was to evaluate the performance of MC in identifying the microbiological content and determining the leukocyte count in synovial fluid (SF) collected from hip and knee revision arthroplasty cases with UNC. It was also assessed whether antibiotic treatment would have been changed if the result from MC had been known.
Method
A total of 89 SF samples from 70 patients (43 female; 27 male) who underwent revision arthroplasty (14 hip; 75 knee) were included. Using European and Bone Joint Infection Society (EBJIS) criteria, 82 cases were classified as infected (77 UNC and 5 septic culture-positive controls), five as non-infected (aseptic culture-negative controls), and two as likely infected, but infected by clinical observation. MC was performed and evaluated together with SF parameters. Antibiotic treatment, clinical outcome, patient demographics and surgical details were analysed.
Results
Overall, 29.1% (23/79) of UNC had a positive yield by MC, of which 2/23 (8.7%) had two microorganisms detected simultaneously. Of the 25 microorganisms identified by MC, 12/25 (48%) were clinically relevant after re-evaluation of the patients’ microbiological history. The microorganisms detected were 5/25 (20%) Streptococcus pneumoniae/mitis, 4/25 (16%) Staphylococcus epidermidis, 3/25 (12%) Cutibacterium acnes, 3/25 (12%) Streptococcus agalactiae, 2/25 (8%) Streptococcus bovis, 2/25 (8%) Staphylococcus aureus, and 2/25 (8%) Haemophilus parainfluenzae. The prevalence of Enterococcus faecalis, Bacteroides fragillis, Staphylococcus lugdunensis, Corynebacterium striatum among all MC results was 1/25 (4%) each species. In total, 13/23 (56%) cases were associated with patients receiving antibiotic therapy at the time of SF collection. The yield for leukocyte counts provided the molecular technique was consistently much higher in the UNC and clearly septic groups than in the clearly aseptic group. Overall, 20/61 (32.8%) patients with UNC could have been managed differently and more accurately after MC assessment.
Conclusions
MC shows clinical value in the diagnosis and management of PJI with UNC. The included leukocyte count shows promising results.
Acknowledgments
This work was partially funded by Inbiome.
