Septic arthritis (SA), also termed infectious arthritis, is a rheumatological emergency characterized by rapid joint destruction and considerable morbidity and mortality (Sharff et al., 2013). Bacterial invasion of the synovial space, most often through hematogenous spread, initiates acute inflammation that can swiftly damage cartilage and subchondral bone (Gigante et al., 2019; Shirtliff and Mader, 2002). SA can occur at any age, but it is most commonly seen in young children and older adults, with a consistent male predominance (Geirsson et al., 2008; Kaandorp et al., 1997a). The reported incidence ranges between 2 and 12 cases per 100 000 annually, with higher rates among individuals with comorbidities such as diabetes mellitus, chronic renal disease, malignancy, or immunosuppressive therapy (Momodu and Savaliya, 2022; Tande and Patel, 2014). Most patients present with acute monoarthritis, typically of the knee, and treatment delays beyond 24–48 h are strongly associated with irreversible joint damage, long-term disability, and increased mortality (Lieber et al., 2018).

Prosthetic joint infection (PJI) represents a particularly severe and complex form of SA, constituting one of the most devastating and costly complications of arthroplasty (Pina et al., 2019; Kurtz et al., 2008). Its incidence is projected to rise in parallel with an aging population, a higher prevalence of comorbid conditions, and the increasing number of joint replacement procedures performed worldwide (Ayoade et al., 2023). Management of PJI is especially difficult because of the ability of microorganisms to form biofilms on prosthetic surfaces. Biofilm-associated infections are protected from host immune responses and antimicrobial activity, often necessitating prolonged or repeated surgical interventions for eradication (Gbejuade et al., 2015).

Coagulase-negative staphylococci (CoNS) are among the leading pathogens implicated in PJI. While frequently regarded as contaminants in native joint infections, CoNS are now recognized as true pathogens in the prosthetic setting due to their biofilm-forming capacity and high rates of methicillin resistance (Tande and Patel, 2014). Contemporary studies confirm that Staphylococcus aureus (S. aureus) and CoNS are the predominant causative agents, although Gram-negative bacilli and enterococci are also encountered with clinically significant frequency (Suardi et al., 2024). Methicillin-resistant S. aureus (MRSA) has been associated with prolonged therapy, increased complications, and higher mortality (Kim et al., 2023). Similarly, the growing prevalence of methicillin-resistant S. epidermidis (MRSE) is of major concern, as these strains are often multi-drug resistant, leaving limited therapeutic options such as glycopeptides and linezolid (Bondarczuk et al., 2017).

Despite advances in diagnostic methods, the accurate identification of joint infections remains a challenge. The diagnosis of PJI, in particular, lacks universally accepted criteria and usually relies on a combination of clinical assessment, laboratory biomarkers, imaging, and microbiological findings. Synovial fluid aspiration and culture continue to be regarded as the diagnostic gold standard, as they enable pathogen identification and antimicrobial susceptibility testing (Parvizi et al., 2011). However, culture sensitivity is often reduced in patients with prior antibiotic exposure or in infections associated with biofilm formation, resulting in false-negative results (Trampuz and Zimmerli, 2006).

Given these challenges, timely and reliable microbiological characterization is crucial for guiding antimicrobial therapy and improving outcomes. Defining the microbial spectrum and resistance patterns in both native SA and PJI is essential not only for diagnostic accuracy but also for informing empirical treatment and stewardship practices. Yet, contemporary large-scale studies directly comparing native and prosthetic joint infections remain limited, particularly in regions with high antimicrobial resistance rates. The present study therefore aimed to investigate the distribution of microorganisms isolated from synovial fluid cultures and to evaluate their antimicrobial resistance profiles. By comparing native and prosthetic joint infections, we sought to highlight etiological differences, support optimization of empirical therapy, and contribute to improved clinical management of both conditions.

SA and PJI are medical emergencies due to their potential for rapid cartilage destruction and long-term disability. Although diagnosis relies on clinical, laboratory, and imaging findings, synovial fluid culture remains the cornerstone despite limited sensitivity and delayed turnaround (Parvizi et al., 2011). In PJI, biofilm formation complicates management and increases the risk of treatment failure (Zimmerli et al., 2004; Gbejuade et al., 2015). Evidence from randomized trials on optimal antibiotic regimens in SA is limited, underscoring the need for strategies informed by local microbiology and resistance trends (Gjika et al., 2019). In this context, our 5-year study analyzed synovial fluid cultures, compared native and prosthetic infections, and described year-to-year changes in antimicrobial resistance patterns.

Overall, 20.5 % of synovial fluid samples yielded microbial growth, lower than positivity rates commonly reported in the literature (42 %–56 %) (Geirsson et al., 2008; Kaandorp et al., 1997a; Ravn et al., 2023; Talsma et al., 2021). This discrepancy likely reflects prior antibiotic exposure, sampling variability, and the absence of adjunctive techniques such as sonication. The low yield emphasizes the continued relevance of culture surveillance and the potential value of complementary molecular diagnostics. Adults accounted for most culture-positive cases (93.1 %), and pediatric cases were comparatively infrequent, consistent with the predominance of PJIs in older populations and the increasing burden of SA with age (Geirsson et al., 2008; Kaandorp et al., 1997b). This distribution likely reflects the inclusion of PJIs, a condition seen primarily in adults. Unlike population-based reports of male predominance in native SA (Geirsson et al., 2008; Mathews et al., 2010), we did not observe a marked sex difference, which may be explained by the inclusion of both native and prosthetic infections, and the demographic profile of arthroplasty recipients (Lenguerrand et al., 2018; Ren et al., 2021; De Mauro et al., 2024).

Positivity rates varied across departments. The highest yields were observed in the Department of Orthopedics and Traumatology (24.2 %) and the Department of Infectious Diseases and Clinical Microbiology (21.6 %), reflecting targeted sampling and frequent intraoperative cultures (Shamdasani et al., 2023). In contrast, lower yields in the Department of Emergency (14.0 %) likely resulted from heterogeneous case presentations and prior empiric therapy (Puzzitiello et al., 2024), whereas the very low rates in Rheumatology (3.6 %), and Physical Medicine and Rehabilitation (2.1 %) mainly reflected cultures obtained to exclude infection in predominantly non-infectious joint diseases (Pal et al., 1997).

A key practical finding was the improved recovery rate when synovial fluid was inoculated into blood culture bottles compared with sterile containers. This observation aligns with prior reports suggesting that higher inoculum volumes and antibiotic-neutralizing media can enhance detection (Cohen et al., 2020; Font-Vizcarra et al., 2010; Li et al., 2019). Together, these findings support routine inoculation of synovial fluid into blood culture bottles as a simple and implementable measure to increase diagnostic sensitivity and potentially reduce false-negative cultures in routine practice.

Nearly half of culture-positive cases were PJIs, characterized by higher age, female predominance, and knee localization, consistent with prior registry data (Nair et al., 2017; Tai et al., 2022; Khakzad et al., 2022). The microbial spectrum was dominated by staphylococci. S. aureus remained the most frequent pathogen overall, confirming its central role across native SA and PJI (Tande and Patel, 2014; Arieli et al., 2021). CoNS were especially frequent in PJIs, underscoring their pathogenic significance in the prosthetic setting through biofilm formation and device-related persistence (Arieli et al., 2021; Chan et al., 2020). Gram-negative bacilli accounted for 22.6 % of isolates, led by P. aeruginosa, K. pneumoniae, and E. coli, again consistent with published series (Tsaras et al., 2012). E. faecalis was isolated in 4.9 % of cases, higher than the 1.9 % reported by Hasbek and Çubuk (2022), and, together with CoNS, was significantly more common in PJI, confirming its strong association with prosthetic material (Renz et al., 2019; Linke et al., 2022). In contrast, S. aureus and Gram-negative bacilli were more evenly distributed between native and prosthetic infections, consistent with prior reports (Linke et al., 2022; Talsma et al., 2021; Gbejuade et al., 2015). Pathogen profiles did not vary by age, consistent with Alexandersson et al. (2024), suggesting that clinical differences across age strata are more likely driven by host factors and comorbidities than by major shifts in microbial etiology.

Year-to-year resistance patterns revealed clinically relevant changes. Methicillin resistance among S. aureus increased over the study period, consistent with broader reports and reinforcing the need to ensure appropriate empirical coverage, particularly given the association between methicillin resistance and adverse outcomes in SA (Ben-Chetrit et al., 2020; Dubost et al., 2014; Kim et al., 2023). Increasing macrolide resistance and inducible clindamycin resistance were also observed, paralleling international findings (Wu et al., 2023). In contrast, glycopeptides, linezolid, and daptomycin maintained full activity, in agreement with European reports (Ritchie et al., 2010; Markwart et al., 2021), supporting their continued role as cornerstone agents for severe Gram-positive joint infections.

Methicillin resistance among CoNS remained consistently high, in line with Turkish and international reports (Hasbek and Çubuk, 2022; Hu et al., 2021). Together with their biofilm capacity, this poses a major therapeutic challenge (Gbejuade et al., 2015). Erythromycin resistance was also high, reaching 72.5 % in 2024, which is consistent with patterns reported elsewhere (Hasbek and Çubuk, 2022; Hu et al., 2021). In contrast, retained susceptibility to glycopeptides and linezolid mirrors previous reports (Hellmark et al., 2009; Suardi et al., 2024) and emphasizes the importance of preserving these last-line options through stewardship. E. faecalis isolates remained universally susceptible in our dataset (Martin et al., 2023); nevertheless, their increased frequency in PJIs warrants clinical vigilance, particularly in mixed infections and in settings with prolonged antimicrobial exposure (Renz et al., 2019).

Resistance patterns among Gram-negative organisms were heterogeneous. For P. aeruginosa, piperacillin–tazobactam resistance decreased across the study years; however, this pattern should be interpreted cautiously, as changes in EUCAST breakpoints may influence apparent year-to-year resistance rates independent of true epidemiological shifts (Ourghanlian et al., 2022; Johnson et al., 2022). Similar breakpoint-related variability has been described for ceftazidime, cefepime, and ciprofloxacin (Wolfensberger et al., 2013). Amikacin resistance remained relatively low, consistent with observations from other regions (Abdeta et al., 2023). These findings highlight the need to interpret longitudinal resistance data in the context of evolving interpretive standards and local stewardship practices.

K. pneumoniae demonstrated substantial β-lactam resistance, and ESBL production increased over the study period (Ramatla et al., 2023). Carbapenem resistance remained clinically important, with variability across years that likely reflect patient factors, antimicrobial use, infection control, and evolving EUCAST breakpoints. These findings parallel European Centre for Disease Prevention and Control (ECDC) surveillance data showing high carbapenem resistance in parts of Europe data (ECDC, 2025).

E. coli also showed high ESBL rates and marked third-generation cephalosporin resistance, whereas carbapenem resistance remained relatively limited, similar to data reported from Pakistan (Ali Khan et al., 2025) but differing from a Romanian PJI cohort reporting the universal susceptibility of E. coli (Dragosloveanu et al., 2025). The high ESBL prevalence in our series most likely reflects regional prescribing practices and ecological pressures.

This study has limitations. The retrospective design may have resulted in incomplete clinical data, including unmeasured prior antibiotic exposure, which could influence culture yield and pathogen recovery. The reliance on conventional culture without adjunctive methods such as sonication or molecular assays may have underestimated the contribution of some organisms. As a single-center study, generalizability to other settings is limited. In addition, year-to-year comparisons of resistance rates may be influenced by changes in EUCAST breakpoints over time. These limitations support the need for prospective multi-center studies incorporating standardized sampling, enhanced diagnostics, and detailed clinical metadata.

In conclusion, our 5-year analysis confirms the predominance of staphylococci in joint infections and highlights the strong association of CoNS and E. faecalis with PJIs. Increasing methicillin resistance among staphylococci and the high prevalence of ESBL-producing Enterobacterales emphasize the importance of careful empirical coverage tailored to local epidemiology; whereas preserved activity of glycopeptides, linezolid, and daptomycin is reassuring for severe Gram-positive infections. Finally, routine inoculation of synovial fluid into blood culture bottles substantially improves diagnostic yield and represents a practical evidence-based strategy to strengthen routine diagnostic workflows and optimize patient management.