Synovial Fluid Sequencing: A Look into the Future of Prosthetic Joint Infection Detection

With the cost of genomic sequencing rapidly decreasing, next‐generation sequencing is being applied on a much larger scale in areas such as clinical diagnostics and infectious disease detection. The purpose of this study was to compare the effectiveness of 16S rRNA (DNA), metagenomic (DNA), and metatranscriptomic (RNA) sequencing of synovial fluid in detecting prosthetic joint infections (PJIs) and to identify potential sources of environmental contamination. Synovial fluid and blood samples were collected from 30 patients receiving a total joint arthroplasty (TJA) at the Rothman Orthopaedic Institute (Philadelphia, PA). Patients undergoing a primary TJA (N=10) were classified as primary while patients undergoing a revision TJA were classified as infected (N=10) if they met the 2018 International Consensus Meeting (ICM) criteria for a PJI or aseptic (N=10) if they did not. Six different negative controls; three during sample collection in the operating room (skin swab, air swab, sterile saline solution) and three during sample processing (extraction blank, no template control, low template control) were collected to assess sources of environmental contamination. Beta diversity analysis of metatranscriptomic (MT) and metagenomic (MG) synovial fluid data showed statistically significant (p=0.001) differential clustering of primary, aseptic, and infected samples, revealing distinct microbial community compositions. Less significant clustering (p > 0.001) between the three cohorts was observed in the 16S rRNA synovial fluid data. Clustering of infected MT samples appeared to be driven by increased counts of Escherichia coli when compared to primary and aseptic samples. This taxon was also identified as the most significantly enriched biomarker (LDA=4.04, p=0.00094) within infected MT samples. Functional gene profiling yielded three significantly expressed genes (LDA>3, p<0.002) in infected samples including FliZ protein and 3‐methyl‐2‐oxobutanoate hydroxymethyltransferase protein coding genes associated with biofilm formation and secondary metabolite synthesis in E. coli , respectively. Concordance of MT and MG data with synovial fluid culture results was 100% (6/6), while 16S rRNA data yielded 0% (0/6) concordance. In negative control samples, the top 10 most abundant bacteria included common contaminants such as Acinetobacter , Bacillus , and Clostridiales but also taxa associated with PJIs including Cutibacterium acnes , Staphylococcus epidermidis , and E. coli . This study illustrates the potential applications of metatranscriptomics to clinical issues such as PJIs. It also underlines the importance of using negative controls to ensure that organisms detected in clinical samples are true pathogens and not contaminants. Future work will focus on exploring the mechanism of initial infection using blood sample data, tracking common contaminants, and better understanding how microbial interactions are driving pathogenicity. Support or Funding Information This work was supported by Contamination Source Identification LLC.