Abstract
Aim
We prospectively evaluated four different microbiological tools for diagnostics of prosthetic joint infections (PJI), and assessed their impact on the categorization of infection according to EBJIS guidelines. We compared culture, in-house real-time mPCR for S. aureus, S. lugdunensis, S. hominis, S. epidermidis, S. capitis, S. haemolyticus, C. acnes (mPCR), broad-spectrum PCR (Molzym) with 16S rRNA V3-V4 amplicon Sanger sequencing (16S PCR), and 16S rRNA V3-V4 amplicon next-generation sequencing (16S NGS) on MiSeq (Ilumina).
Methods
A total of 341 samples (sonication fluid, tissue biopsy, synovial fluid) were collected from 32 patients with suspected PJI who underwent 56 revision surgeries at the Orthopaedic Centre University Hospital Ljubljana, between 2022 and 2024. Samples were processed using standard protocols for routine culture, followed by DNA isolation using the MagnaPure24 (Roche). All samples were tested with mPCR, and an additional ≥4 samples from each revision (244 in total) were subjected to further metagenomic analysis. Culture results were considered positive if the same microorganism was detected in ≥2 samples, ≥50 CFU/ml were present in the sonication fluid, or ≥1 sample was positive for a more virulent microorganism or if the patient had received antibiotic treatment.
Results
Each tool demonstrated high sensitivity for correct EBJIS categorization (100% culture and 16S NGS, 96.88% mPCR and 16S PCR). The highest specificity was observed with mPCR and 16S PCR (87.5%), while culture (79.17%) and NGS (37.5%) showed lower specificity. In 27% (15/56) of revisions, all microbiological tests were negative, although infection was confirmed with histology in one case, and four cases were classified as infection-likely based on clinical signs. In 20% (11/56) of cases, all microbiological tests were positive; in three cases a combination of other EBJIS criteria (without microbiology) categorized the episodes as infection-likely and one as infection-unlikely, emphasizing the importance of microbiological tests in diagnostic criteria. In 43% (24/56) of revisions categorized as infection-unlikely using a combination of other EBJIS criteria, five had positive culture, and three had positive mPCR and 16S PCR. Fifteen (62%) had positive 16S NGS, 12 due to a low number of reads, which may indicate low-grade infection or possible contamination.
Conclusion
To date, no test can be established as the ultimate gold standard. The lack of interpretation criteria can result in low specificity of some methods, as the threshold is difficult to determine. A multidisciplinary approach with combination of microbiological tools is still considered the most efficient.
