Development and evaluation of rapid methods for diagnosis of prosthetic joint infection

Prosthetic joint infection (PJI) is a devastating complication of total joint replacement resulting in severe pain, functional impairment and high mortality. While a small minority of joint replacements become infected, accurate diagnosis and appropriate management are essential to restore function and prevent excess morbidity. The gold standard for diagnosis, microbial culture, is slow and insensitive, hence other tests (e.g. inflammatory markers) are often used to get a more accurate diagnostic. The inability to differentiate between low-grade infections and aseptic loosening leads to some patients undergoing numerous investigations and unnecessary two-stage revisions, with higher cost and patient morbidity. Rapid lateral flow tests have the potential for early rule-out/in of infection, directing further diagnostics and appropriate patient management. We tested a commercially available calprotectin lateral flow assay on 69 synovial fluid samples from patients undergoing joint revision at Norfolk and Norwich University Hospital (NNUH). The test sensitivity and specificity against the International Consensus Meeting (ICM) criteria for infection were 75% and 75.56%, respectively. Adjusting results following a clinical review of patient outcomes, the sensitivity improved to 94.74% and specificity was 78%. The low specificity was due to false positive results caused by metallosis or severe osteolysis. The assay showed a high negative predictive value of 96.50% compared to clinical review which would make this a good rule-out test for PJI. Clinical metagenomics (CMg) has the potential to provide faster and more comprehensive results than culture, capable of detecting any pathogen (bacteria, viruses and fungi) within hours in a single test. The main challenge in achieving rapid turnaround CMg has been the high ratio of human:pathogen DNA present in clinical samples, resulting in slow and expensive sequencing tests. We developed and optimised a rapid CMg pipeline for the diagnosis of PJI that overcomes these challenges. The pipeline includes: a simple and highly efficient differential-lysis-based host DNA depletion step, automated DNA extraction, rapid library preparation, low-input nanopore sequencing and real-time identification of microorganisms using Epi2Me (Oxford Nanopore). The pipeline was optimised using excess sonication fluid samples from suspected PJI patients at the NNUH. The optimised pipeline was then evaluated in a study using MicroDTTect whole prosthesis sampling from 42 revision patients. The developed CMg test had a turnaround time of six hours for the identification of bacterial pathogens with 88.89% specificity and 60% sensitivity compared to routine culture. Sensitivity increased to 92% after investigation into discordant samples as many of the pathogens missed by metagenomics were likely to be culture contaminants. The performance was also better when compared to culture of the MicroDTTect fluid; 73.33% sensitivity and 85.19% specificity. One of the major challenges in diagnosing PJI is determining whether detection of skin commensal bacteria (which can cause PJI) is indicative of infection or contamination of the sample by skin flora during sampling. We set out to identify novel biomarkers of biofilm formation, which if detected in an organism, would indicate that it had recently been in a biofilm and had originated from the PJI and not from the skin. Staphylococcal strains were transformed using electroporation and plasmids specifically designed for Transposon-Insertion Sequencing (TIS) to develop a Staphylococcus caprae library of ~250,000 insertion mutants. This library could be used for gene function profiling to discover new biomarkers of biofilm formation, which could eventually be translated into novel diagnostic tests. In conclusion, we investigated three important areas in PJI: detecting host biomarkers, utilising CMg for the rapid identification of pathogens and antimicrobial resistance (AMR) and biofilm biomarker discovery through TIS. We think these three approaches could be combined into a single test in the future, sequencing bacterial biofilm and host mRNA biomarkers and identifying bacteria and AMR. This test would rapidly and accurately detect PJI in patients, improving patient management, reducing hospital costs and reducing both patient morbidity and mortality