gms | German Medical Science

Objectives: Joint infection is challenging, as it increases morbidity and mortality rates in patients with prosthetic joints and raises the risk for secondary osteoarthritis in native joints. Accordingly, adequate initial antibiotic therapy targeting the predominant bacteria is crucial. The causative pathogen spectrum depends on the infection site, comorbidities, and other determinants. Nevertheless, it is arguable how the presence of an endoprosthesis influences the type of detectable bacteria. To this end, we examined differences in the spectrum of pathogens based on synovial fluid analyses in patients with periprosthetic joint infections (PJI) compared to patients with infections of native joints (NJI).

Methods: We retrospectively evaluated microbiologically positive synovial fluid samples of 192 consecutive patients obtained from January 2018 to January 2020 in a tertiary care university hospital. In these samples, we examined patient characteristics and bacteriological differentiation. We observed patients separately, depending on the presence or absence of a prosthesis or implant in the infected joint. Mann-Whitney-U was applied for intergroup comparison of metrically distributed parameters. For nominal data, crosstabs and chi-squared tests were implemented using the contingency coefficient (C) to confirm significant differences. The confidence interval was set to 95%.

Results and Conclusion: Overall, 132 patients suffered from PJI and 60 patients showed NJIs. The average patient age was significantly higher in the PJI group compared to the NJI group (70.8 vs. 56.3 years, p<0.05). Common localizations of infections were the hip (50%), followed by the knee (39.5%) and the shoulder (3.6%). Concerning the hip joint, PJIs were 3.9-fold more frequent than NJIs (65.2% vs. 16.7%, p<0.05). The most commonly isolated pathogen types were coagulase-negative staphylococci (CNS, 28%), followed by staphylococcus aureus (STAU, 26.7%), and other bacteria (26.3%). We observed a significant dependence between the pathogen type and the presence of an endoprosthesis (p<0.05). Accordingly, detection rates of CNS occurred more frequently in PJIs compared to NJIs (33.9% vs. 13.4%). In contrast, STAU was observed more than 3-fold times more often in the NJI compared with the PJI group (52.2% vs. 16.4%, p<0.05).

The significant differences in the pathogen spectrum of PJIs compared to NJIs may necessitate a differentiated consideration of early empiric antibiotic regimes dependent on the presence or absence of a joint prosthesis.