Periprosthetic joint infection (PJI) is one of the most significant complications after total hip and knee arthroplasty and remains a leading cause of failure and revision surgery. Although its incidence after primary arthroplasty is relatively low, with an incidence ranging from approximately 1.5 % to 2.5 % of cases, the growing number of primary and revision procedures means that the absolute burden of PJI is steadily increasing (Flurin et al., 2019). These infections lead to repeated operations, prolonged antibiotic therapy, impaired function, and a substantial consumption of healthcare resources (Marculescu and Cantey, 2008; Tan et al., 2016; Wimmer et al., 2016).

Staphylococci are the predominant pathogens in PJI, and, within this group, coagulase-negative staphylococci (CoNS), especially Staphylococcus epidermidis, have emerged as key etiological agents in hip and knee PJI, accounting for up to 23 %–25 % of the cases in several studies (Lange et al., 2025). Traditionally regarded to be low-virulence commensals, S. epidermidis strains have adapted remarkably well to the prosthetic joint environment. Their ability to adhere to biomaterials and to form biofilm favors persistence on the implant surface and substantially compromises the effectiveness of systemic antimicrobial therapy (Rohde et al., 2010).

When treating PJI it is important to consider planktonic bacteria, especially relevant during the initial phases of acute infections (Ariza et al., 2017). Initial therapy should target rapidly growing planktonic bacteria, typically with β-lactams, lipopeptides, or glycopeptides, followed by orally bioavailable agents with good bone penetration combined with drugs active against biofilm-embedded bacteria, such as rifampin (Auñón et al., 2022). Accordingly, characterization of antimicrobial susceptibility – particularly to β-lactams, fluoroquinolones, and rifampin – is essential in S. epidermidis PJI as resistance to these agents significantly impacts treatment strategy and outcomes.

Methicillin resistance is common in S. epidermidis PJI, reflecting hospital antibiotic pressure and the organism's capacity to accumulate resistance determinants (reported in 18 %–45 % of isolates) (Jiang et al., 2025; Lange et al., 2025). Methicillin resistance (MRSE) limits β-lactam options and may affect early management, while fluoroquinolone and rifampin resistance further compromise standard oral and biofilm-active regimens (Hischebeth et al., 2019).

A similar situation has been described for fluoroquinolone resistance. Several studies have identified the absence of an active fluoroquinolone-based regimen as a factor associated with treatment failure in Staphylococcus aureus prosthetic joint infection; nonetheless, these analyses were not specifically focused on coagulase-negative staphylococci, and results were reported for S. aureus rather than for CoNS species as a distinct group (Lora-Tamayo et al., 2013). Several authors have reported poorer outcomes in staphylococcal prosthetic joint infection caused by rifampin-resistant isolates. In the landmark study by Zimmerli et al. (1998), which included S. aureus and other staphylococcal species, the absence of rifampin activity was associated with significantly higher rates of treatment failure, establishing rifampin as a major determinant of prognosis in implant-associated staphylococcal infections. Similarly, Tornero et al. (2014) analyzed risk factors for failure in early prosthetic joint infection managed with debridement, antibiotics, and implant retention, identifying rifampin resistance as the main treatment-related factor associated with failure among gram-positive microorganisms overall. However, again, these analyses did not specifically focus on CoNS.

The aim of the present study is to characterize prosthetic joint infections caused by S. epidermidis at our institution by describing their epidemiology, clinical presentation, microbiological features, surgical and antimicrobial management, and outcomes.

A total of 615 patients diagnosed with periprosthetic joint infection (PJI) were identified and initially considered for inclusion in the present study. After applying exclusion criteria, 105 cases (20.35 %) were caused by S. epidermidis.

The mean age of the cohort was 71 years old (SD: 10.4). Females accounted for 56 % of patients. The knee joint was the most frequently involved site, accounting for 51 % of PJIs, while the hip was affected in 49 % of cases.

When considering the type of implant involved, 59 % of infections occurred in primary prosthetic implants, and 41 % occurred in revision arthroplasties. A total of 83 patients (79 %) met the chronic infection definition, whereas 22 (20 %) were diagnosed as acute infection.

From a surgical risk perspective, the mean ASA Score was 2.49; furthermore, the mean CCI score was 3.73. Assessment of BMI showed that 45.7 % of the study population suffer from obesity (BMI ≥ 30 kg m⁻²).

As for related risk factors, 14.3 % of the patients had diabetes mellitus. Demographic data and comorbidity profiles are shown in Table 1.

In terms of surgery, 29 patients underwent DAIR (29/105, 27.6 %), 30 underwent one-stage revision (30/105, 28.6 %), 41 underwent two-stage revision (41/105, 39.0 %), and 5 underwent other procedures (5/105, 4.8 %) such as resection arthroplasty or arthrodesis. Surgical data are shown in Table 2 for further reference.

Overall, treatment was successful in 82 patients (82/105, 78.1 %). Between the surgical procedures, the success rates were 75.9 % for DAIR, 93.3 % for one-stage revision, and 73.1 % for two-stage revision. None of the recorded demographic characteristics or risk factors were significantly associated with treatment failure. Similarly, neither MDR/XDR infection nor rifampin resistance was significantly associated with any of the collected baseline variables. Of the 23 failures, 10 (43.5 %) were due to persistent infection, and 9 (39.1 %) were due to superinfection. Superinfections were most frequently caused by Enterobacteriaceae (44.5 %), followed by Proteus mirabilis (33.3 %), Klebsiella pneumoniae (11.1 %), and Candida albicans (11.1 %).

The mean duration of antimicrobial therapy was 10.1 weeks overall. Patients treated with DAIR received longer courses (mean: 11.8 weeks) compared with those undergoing one-stage revision (10.6 weeks). In contrast, patients managed with two-stage revision had a shorter mean treatment duration (8.5 weeks).

In our single-center cohort of 105 monomicrobial S. epidermidis PJIs, the overall treatment success rate was 78.1 %. The most clinically relevant findings were that rifampin resistance and MDR/XDR phenotypes were each associated with significantly lower success, whereas methicillin resistance and fluoroquinolone resistance were not statistically associated with failure in this series.

The proportion of PJIs caused by S. epidermidis (20.3 %) in our cohort is consistent with contemporary epidemiology, in which CoNS and S. epidermidis represent a major fraction of causative organisms. Large institutional series and reviews commonly report S. epidermidis rates around the mid-20 % range among PJI pathogens, supporting the fact that our microbiological distribution is representative of current practice (Jiang et al., 2025; Lange et al., 2025)

A high prevalence of resistant pathogens was observed at 72.3 % MRSE, 53.3 % fluoroquinolone-resistant, 24.8 % rifampin-resistant, and 43.8 % MDR/XDR, reflecting the growing prevalence of multidrug resistance in PJI (Lange et al., 2025).

We observed the highest success with one-stage revision (93.3 %), followed by DAIR (75.9 %) and two-stage revision (73.1 %). This could likely reflect a selection bias effect rather than intrinsic superiority of one-stage exchange. In many centers, two-stage revision is preferentially used for more complex infections, which can depress success rates. However, the favorable results of the one-stage approach in our cohort are consistent with European practice, where careful patient selection and a feasible targeted antimicrobial plan support its effectiveness (Lange et al., 2025).

Conversely, implant retention strategies can perform poorly in CoNS infections, particularly in unfavorable microbiological settings. In one cohort of CoNS knee PJI, the overall 1-year success rate was 47 %, and outcomes after DAIR were especially low (Charalambous et al., 2022).  The relatively favorable DAIR success observed in our cohort should be interpreted cautiously. DAIR outcomes depend heavily on timing, implant stability, adequacy of debridement, exchange of modular components, host factors, and – especially in staphylococcal infections – the availability of an active biofilm-oriented regimen. Available evidence suggests that patients in whom rifampin can be used tend to achieve higher DAIR success rates, with cure rates around 70 % compared to approximately 50 % in those not receiving rifampin (Scheper et al., 2021).

Our findings support the fact that, in S. epidermidis PJI, especially when resistance limits effective antibiotic regimens, adequate source control through implant removal may be critical to achieve infection eradication.

Successful management of staphylococcal PJI requires effective anti-biofilm combination therapy beyond activity against planktonic bacteria. The randomized trial by Zimmerli et al. demonstrated superior outcomes with rifampin-based regimens in implant-related infections treated with debridement (Zimmerli et al., 1998). This benefit is reflected in guideline recommendations. SEIMC guidance highlights rifampin as the key agent against biofilm-associated staphylococci, particularly in DAIR, and recommends rifampin–fluoroquinolone combinations as a preferred oral regimen when feasible (Lora-Tamayo et al., 2013).

Within this framework, our finding that rifampin-resistant S. epidermidis had significantly worse outcomes is both clinically coherent and relevant. Rifampin resistance eliminates one of the few consistently supported oral anti-biofilm strategies and forces reliance on rifampin-sparing regimens, for which evidence is less robust and tolerability can be limiting (Lora-Tamayo et al., 2013; Zimmerli et al., 1998). Where rifampin cannot be used, alternative combinations (e.g., tetracycline-based strategies such as minocycline-containing regimens) have been proposed, but supporting data are comparatively limited and often non-randomized (Bart et al., 2020).

MDR/XDR phenotypes were also associated with failure, likely through multiple mechanisms. These profiles limit the feasibility of constructing effective, biofilm-active, and well-tolerated regimens and may reflect prior antibiotic exposure or more complex clinical scenarios. Moreover, MDR frequently coexists with rifampin and fluoroquinolone resistance, further restricting optimal consolidation strategies (Magiorakos et al., 2012). In our series, resistance to rifampin and levofloxacin in combination showed the strongest association with treatment failure, with an odds ratio of approximately 8.

In contrast to rifampin resistance and MDR/XDR, methicillin resistance was not associated with worse outcomes in our cohort, despite reports from two-stage revision series showing lower eradication rates in MRSE compared with methicillin-susceptible infections (Hischebeth et al., 2019). Similarly, two-stage revision for methicillin-resistant hip PJI has been reported, with infection control rates around 71 % (methicillin-resistant) vs. ∼ 90 % (controls), highlighting the potential adverse impact of resistant organisms, even in exchange settings (Santoso et al., 2020). This discrepancy likely reflects differences in terms of case mix and surgical approach rather than the absence of a biological disadvantage for MRSE as methicillin resistance alone may be less impactful when adequate source control and effective therapy are achieved. Fluoroquinolone resistance showed a trend toward lower success, although without statistical significance, possibly due to limited power or alternative regimens. However, because rifampin-based strategies commonly rely on a fluoroquinolone partner, this resistance pattern remains clinically relevant and may have a greater prognostic impact in larger or DAIR-focused cohorts (Lora-Tamayo et al., 2013; Zimmerli et al., 1998).

Overall, our findings suggest that outcomes in S. epidermidis PJI depend largely on the feasibility of administering an effective biofilm-active regimen, often determined by rifampin susceptibility. In the presence of rifampin resistance or MDR/XDR profiles, implant exchange may be preferable to DAIR, given the greater reliance of retention strategies on optimal antimicrobial therapy (Charalambous et al., 2022; Lora-Tamayo et al., 2013, 2013).

The practical implication of our findings is that antimicrobial resistance should influence not only antibiotic selection but also the surgical strategy. In susceptible isolates, particularly when rifampin and a suitable partner drug remain available, implant-retention strategies may still be reasonable in selected acute cases. By contrast, when rifampin resistance or combined rifampin–fluoroquinolone resistance is present, the feasibility of a standard oral biofilm-active regimen is substantially reduced. In those scenarios, surgeons and infection specialists should have a lower threshold for implant exchange, particularly when other unfavorable features coexist, such as chronic infection, compromised soft tissues, multiple prior procedures, or limited modular exchange. In this setting, the use of new antimicrobials with high antibiofilm activity (like linezolid, tedizolid, dalbavancin, and others) can be used according to the individualized antimicrobial susceptibility of each strain and the specific characteristics of the patients (Valour et al., 2025)

Finally, emerging anti-biofilm strategies could become particularly relevant in resistant S. epidermidis PJI, where traditional treatment options are often limited. There is increasing focus on approaches such as bacteriophage therapy, phage–antibiotic combinations, biofilm-disrupting agents, and alternative drugs with potential anti-biofilm activity. However, the evidence remains limited and is mostly based on experimental ground, small series, or selected compassionate-use cases. For now, these strategies should be used as adjuncts rather than as a replacement for established surgical and antimicrobial approaches (Ortiz-Cartagena et al., 2025).

Despite being one of the largest single-center series dedicated to monomicrobial S. epidermidis PJI, this study's retrospective design and long study period introduce potential confounding, and surgical comparisons are subject to selection bias. Additionally, resistance profiles were used as surrogates for treatment feasibility, and detailed regimen-level analyses would better clarify their impact on outcomes.

In this cohort of monomicrobial S. epidermidis PJI, rifampin resistance and MDR/XDR phenotypes were associated with significantly poorer outcomes, reinforcing the clinical importance of maintaining access to biofilm-active combination therapy and of aligning surgical strategy with microbiological “treatability”. The strong performance of one-stage exchange in selected patients suggests that, when source control is adequate and antimicrobial options are favorable, high success is achievable, even in a population with a substantial antimicrobial resistance pattern.