Prosthetic joint infection (PJI) is a severe complication of knee and hip arthroplasty, and a leading cause of revision surgery (De Ab Cobra et al., 2022; Khan et al., 2016; Koh et al., 2017). Early detection and precise diagnosis are critical, as PJI impairs function and often necessitates invasive treatment with a high risk of adverse events. However, no universally accepted definition exists for PJIs in clinical practice (Mcnally et al., 2021). This lack of consensus stems from factors including the condition’s complexity, geographical variations in clinical practice, reliance on expensive diagnostic tests, and differing opinions on test accuracy (Villa et al., 2020). Laboratory markers, such as white blood cell counts, polymorphonuclear percentage, C-reactive protein, and erythrocyte sedimentation rate, have been considered highly specific; however, recent studies have shown that cut-off values can vary based on the causative organism (Kheir et al., 2018). Serum markers such as leukocyte esterase, interleukin (IL)-6, IL-8, and alpha-defensin show good diagnostic utility but are prone to false-positive conditions such as metallosis, gout, and other inflammatory diseases (Lee et al., 2017; Saleh et al., 2017; Mcnally et al., 2021). Additionally, synovial culture remains necessary to identify causative microorganisms (Carli et al., 2019). Consequently, diagnosing PJI requires an integrated approach that incorporates clinical assessment, serum and synovial fluid biomarkers, histopathological examination, and microbiological culture (Kheir et al., 2018; De Ab Cobra et al., 2022; Mcnally et al., 2021).

Cell-free deoxyribonucleic acid (cfDNA) comprises extracellular DNA fragments detectable in nearly all biological fluids, including blood, cerebrospinal fluid, urine, and synovial fluid (Chan et al., 2003; De Ab Cobra et al., 2022; Peng et al., 2017). Circulating cfDNA originates from decomposing host cells and microbial invaders that release nucleic acids into body fluids (Jahr et al., 2001). Conventional culture-based methods have limitations, including prolonged turnaround times and difficulty in identifying non-culturable pathogens (Han et al., 2020). Consequently, cfDNA-based pathogen detection is beneficial for identifying various pathogens in PJI and optimizing antibiotic selection. The study evaluated the potential of synovial cfDNA metagenomic analysis for PJI detection and microbial identification, with implications for improving PJI diagnosis and treatment following joint surgery.

This study confirmed that cfDNA concentrations in synovial fluid are elevated in patients with PJI and explored the use of cfDNA metagenomic analysis for pathogen identification. PJI is a severe complication following total hip and knee arthroplasty (Ting and Della Valle, 2017), with an estimated economic burden of $60,000–$100,000 per patient and significant morbidity and mortality (Bozic and Ries, 2005; Kurtz et al., 2007; Lavernia et al., 2006). Despite their significant clinical and economic impact, standardized definitions and diagnostic protocols for PJI are lacking (Ting and Della Valle, 2017). Our findings suggest that cfDNA technology may facilitate more rapid and accurate PJI diagnosis. cfDNA-based approaches have been successfully employed in non-invasive prenatal testing, organ transplant rejection monitoring, and liquid biopsy in oncology (Camargo et al., 2019; Bloom et al., 2017; Schütz et al., 2017; De Vlaminck et al., 2014; Allyse and Wick, 2018; Rossi et al., 2018). Recently, this technology has been adapted to diagnose infectious diseases (De Vlaminck et al., 2015; Hong et al., 2018; Lee et al., 2023). In PJI, elevated cfDNA concentrations may enable earlier disease detection, expedite treatment initiation, and improve patient outcomes. This approach could also reduce unnecessary antibiotic use, hospital admissions, and length of hospital stay.

The observed increase in cfDNA concentrations in PJI compared with non-PJI cases was consistent with previous findings. Lögters et al. demonstrated that cfDNA elevation in infections is primarily driven by cell death, particularly through neutrophil extracellular trap (NET) release (Lögters et al., 2009). In a study by Cobra et al., cfDNA concentrations were elevated in the synovial fluid of patients with PJI; however, the source of cfDNA remained unclear, with infecting bacteria proposed as a potential origin (De Ab Cobra et al., 2022). Our study confirmed increased cfDNA concentrations in the synovial fluid of patients with PJI and demonstrated through metagenomic analysis that cfDNA originates from NETosis and bacterial sources.

Rapid diagnosis of PJI and identification of the causative organism are crucial. cfDNA metagenomic analysis detects pathogens at the genus or species level and may provide results faster than traditional culture-based testing (Fung et al., 2017; Hogan et al., 2021). In this study, cfDNA sequencing successfully identified pathogens in three cases where synovial culture failed. Notably, two of these cases involved a history of antibiotic administration within two weeks prior to synovial aspiration. cfDNA sequencing has the advantage of identifying pathogens without being affected by prior antibiotic exposure, a benefit that has been demonstrated in previous studies (Wang et al., 2021). Unlike traditional culture-based methods, which may fail to detect pathogens if antibiotics suppress bacterial growth, cfDNA sequencing detects fragments of microbial DNA present in the bloodstream or infected tissues regardless of the viability of the organism. This advantage makes cfDNA a valuable tool for diagnosing infections, particularly in cases where antibiotics have already been administered.

In comparison with other molecular diagnostic techniques, cfDNA sequencing offers distinct advantages and limitations. Multiplex PCR is rapid but limited to predefined panels, while shotgun metagenomic sequencing (mNGS) provides broader genomic information, including resistance determinants, but is often hampered by high host DNA background and higher costs. cfDNA analysis enriches for extracellular fragments, potentially reducing host interference and enhancing sensitivity, though it cannot yet replace mNGS for resistance profiling. Notably, in two cases, cfDNA sequencing identified multiple microorganisms when synovial culture detected only one, underscoring its potential value in polymicrobial infections.

Sonication culture shows good diagnostic value among prior culture methods compared with conventional bacterial culture with higher sensitivity (Portillo et al., 2014; Liu et al., 2017). In our study, sonication culture identified pathogens that cfDNA analysis did not detect in three cases. This finding suggests that bacteria embedded within biofilms on implant devices may be difficult to detect through cfDNA analysis of the synovial fluid. However, sonication culture requires implant removal, limiting its applicability to post-surgical cases.

Antibiotic selection was based on conventional culture methods (synovial or blood), sonication, or tissue culture, as cfDNA results were not available to clinicians during the study period. However, if cfDNA microorganism results are shared with clinicians, treatment using conventional methods may be enhanced, especially in culture-negative cases. Microbial cfDNA results can help reduce the use of unnecessary antibiotics, leading to antibiotic resistance and side effects. Incorporation of cfDNA could enhance antibiotic stewardship (Yu et al., 2021). Compared with conventional cultures, which typically require 3–7 days for results and are relatively inexpensive, cfDNA sequencing has a higher per-test cost but offers results within 24–48 hours and broader pathogen detection. This trade-off appears reasonable, given the potential to reduce unnecessary antibiotic use and hospital stay through earlier diagnosis.

The possibility of contamination from skin commensals and environmental organisms remains a challenge in cfDNA sequencing. Although we attempted to exclude contaminants by identifying consistent patterns across samples, we acknowledge that whole-genome sequencing (WGS) with SNP analysis would more rigorously confirm whether identical strains are detected by both cfDNA and culture. Due to funding limitations, this was not performed in the present study but will be considered in future investigations.

The primary limitations of this study include its small sample size and single-center design, which may limit the generalizability of the findings. Given the etiological heterogeneity of PJI, a larger, multicenter cohort will be required to validate these results and to improve the reliability of cfDNA-based diagnostic approaches. Additionally, the short read length of the sequencing technology complicates the identification of Staphylococcus species. Larger, multicenter studies are required to validate these findings and provide more robust evidence on the diagnostic accuracy of culture methods and cfDNA analysis in PJIs. Another practical limitation is that synovial fluid is not always available in patients with suspected PJI, particularly after synovectomy. While we attempted to include other sample types, the limited number of cases precluded meaningful conclusions. Future studies should incorporate additional specimen types to broaden the applicability of cfDNA testing.

In summary, cfDNA concentrations were significantly higher in patients with PJI, and cfDNA sequencing complemented culture by detecting pathogens in antibiotic-exposed and polymicrobial cases. These findings support cfDNA as a valuable adjunctive tool for improving the diagnostic workflow of PJI.