The diagnosis of periprosthetic joint infection (PJI) is multifaceted and guided by consensus definitions from specialist societies (Parvizi et al., 2011; Osmon et al., 2013; Parvizi et al., 2014, 2018; McNally et al., 2021). These definitions incorporate a range of diagnostic modalities, including clinical findings, serum and synovial biomarkers, microbiological cultures, and histopathological evaluation. Among the clinical features, a sinus tract communicating with the prosthesis and visible purulence surrounding the prosthesis during revision surgery are most prominent. While a sinus tract is universally regarded definitive evidence of infection according to all PJI definitions (McNally et al., 2021; Parvizi et al., 2018; Osmon et al., 2013; Parvizi et al., 2011, 2014), purulence should be interpreted with greater caution. The IDSA definition considers purulence a major (confirmatory) criterion (Osmon et al., 2013). In contrast, the MSIS 2011 definition included purulence as a minor criterion (Parvizi et al., 2011), but it was excluded in the revised MSIS/ICM 2013 definition (Parvizi et al., 2014). In the ICM 2018 and EBJIS 2021 definitions, purulence was again included as a minor or suggestive criterion (McNally et al., 2021; Parvizi et al., 2018). The reason for this debate is likely that some aseptic conditions, including adverse local tissue reactions (ALTRs, frequently seen in metal-on-metal articulations, trunnionosis, or other instances with dynamic contact of metallic components) and crystal arthropathy, can present with macroscopically cloudy or turbid fluid (Trebse and Roskar, 2021; McNally et al., 2021; Alijanipour et al., 2015), making it difficult to distinguish these cases from true PJI with purulence.

However, although sinus tract and purulence are generally well recognized in clinical practice as possible indicators of PJI, evidence-based literature on their diagnostic value is still lacking. In particular, studies focusing on intraoperative purulence are scarce and often unsystematic. and lack standardized definitions. Furthermore, robust data quantifying the diagnostic accuracy of both sinus tract and purulence are currently unavailable, making it difficult for clinicians to reliably interpret these findings in routine practice. Given the clinical relevance of these signs, a detailed evaluation of their diagnostic performance is warranted.

Therefore, the aim of this study was to assess the diagnostic accuracy of the clinical signs of sinus tract and purulence surrounding the prosthesis for the confirmation of PJI in patients undergoing revision hip or knee arthroplasty. In addition, the microbial profile and patient outcomes associated with these clinical signs were determined.

In our study, the presence of a sinus tract communicating with the prosthesis or joint demonstrated an ideal specificity of 100 % (95 % CI: (98.3–100)) but a low sensitivity of 12.7 % (8.8–17.5; sinus tract). Although included in all infection definitions as a major or confirmatory criterion for PJI (McNally et al., 2021; Osmon et al., 2013; Parvizi et al., 2011, 2014, 2018), to our knowledge, no evidence to date has reported on the diagnostic value of a sinus tract. Based on our results, this clinical sign confirms infection, but its absence does not exclude PJI, as has been previously recognized but never accurately analysed. Accordingly, the presence of a sinus tract can therefore continue to be recommended as a confirmatory criterion for diagnosing PJI.

However, the definition of a sinus tract remains an important yet debated aspect in PJI diagnosis. Some consider prolonged postoperative wound drainage as a sinus tract, whereas others restrict the term to a chronic infection characterized by an epithelium-lined channel connecting the prosthesis to the skin. A clear distinction between transient wound leakage and a true sinus tract with a fully developed epithelialized wall is essential. Infection-induced inflammation can lead to accumulation of synovial fluid within the joint. When intra-articular pressure increases, this fluid may follow the path of least resistance to the skin, forming an immature tract. Persistent infection with ongoing fluid production can promote epithelialization, resulting in a mature sinus tract capable of continuous drainage. A study on chronic sinuses as therapy for treatment-resistant PJIs has shown that creating an iatrogenic tract with a large-diameter drainage tube requires several weeks (6 weeks) for stabilization and epithelialization (Klim et al., 2023), emphasizing that a mature sinus tract develops only over a prolonged time. In chronic infections, the slow and persistent accumulation of fluid under low pressure allows epithelialization and biofilm maturation, marking a true sign of chronic infection. In the present cohort, sinus tracts were observed exclusively in chronic infections. In contrast, skin breakdown or short-term wound leakage in the early postoperative period typically represents immature or incomplete tracts (resulting from the surgical procedure and wound closure) that may reopen under minimal pressure and do not indicate a fully established sinus tract or mature biofilm. In addition, non-infectious factors, such as anticoagulation therapy or malnutrition, can also contribute to postoperative wound drainage (Chotanaphuti et al., 2019). Therefore, a true sinus tract should be strictly defined as a chronic, epithelium-lined channel connecting the prosthesis or joint to the skin, as applied in this study. Nevertheless, we acknowledge that the clinical distinction between prolonged postoperative wound drainage and early sinus tract formation remains challenging and is still subject to ongoing debate. Therefore, patients with early postoperative infections were excluded from the present study in order to minimize potential bias related to postoperative wound drainage.

Importantly, sinus tract formation can occur with any pathogen, depending on microbial virulence, the host immune response, and the surrounding soft tissue envelope. However, certain pathogens were more frequently associated with sinus tract development. In this cohort, the most common pathogens leading to sinus tract formation were Candida albicans (50 %), polymicrobial infections (32 %), Cutibacterium spp. (30.0 %), and Staphylococcus aureus (13 %). Higher prevalences have been reported in previous studies by Luo et al. (2025) and Budin et al. (2024), who analysed the microbial distribution in knee and hip PJI. In knee PJI, polymicrobial infections (42 %), Staphylococcus aureus (24 %), and Enterococcus faecalis (13 %) were most common (Luo et al., 2025), whereas, in hip PJI, polymicrobial (48 %), Staphylococcus aureus (31 %), and Staphylococcus epidermidis (27 %) predominated (Budin et al., 2024). Moreover, polymicrobial infections and Gram-negative bacteria have been previously linked to sinus tract formation (Tan et al., 2016; Xu et al., 2019). In our cohort, however, the number of Enterococcus spp. and Gram-negative infections was small (n=7), precluding meaningful comparisons for these pathogens.

In previous studies, the presence of a sinus has been a predictor of treatment failure (Kheir et al., 2018; Xu et al., 2024, 2019), which is in line with our findings. Patients with a sinus tract formation showed significantly lower survival rates (49.8 % vs. 81.3 % at 36 months, p=0.011) compared with patients without. This striking difference underscores the challenges of managing infections with compromised soft tissue, regardless of whether a one-stage or two-stage revision is performed (Thakrar et al., 2019; Xu et al., 2019). Adequate soft tissue coverage, often with a muscle flap, is frequently required to achieve infection eradication. In our cohort, four patients (n=4/31, 13 %) with knee PJI underwent reconstruction using a local medial gastrocnemius flap. Three proceeded to reimplantation; one later required amputation, while the remaining two remained free of reinfection at 2-year follow-up. These cases illustrate the complexity and difficulty of managing PJI with poor soft tissue conditions, particularly around the knee.

Regarding purulence surrounding the prosthesis, an excellent specificity of 100 %, comparable to that of sinus tract, was observed in our study. Sensitivity remained higher but still insufficient at 35 %. Overall, evidence on the diagnostic accuracy of intraoperative purulence remains limited. Shohat et. al. identified purulence as an important intraoperative marker for chronic PJI in their logistic regression model (Shohat et al., 2019). In contrast, Alijanipour et. al. reported a sensitivity of 82 % but a specificity of only 32 % based on the ICM definition of 2013 (Parvizi et al., 2014). Although the authors acknowledged the limitations of purulence, they did not exclude confounding conditions (such as ALTR, inflammatory arthopathy). Furthermore, their definition of purulence differed from ours as pus was assumed to be present when the surgical findings appeared to be compatible with infection in their study (Alijanipour et al., 2015). By excluding confounding conditions and relying solely on the surgeon's explicit documentation of purulence in the operative report, we observed no false-positive cases in our cohort, aligning with a review on the interpretation of PJI diagnostic tests (Trebse and Roskar, 2021). The authors concluded that, once confounders are excluded, purulence is no longer a controversial criterion and approaches a specificity of 100 %. This underlines its value as an indicator of PJI, particularly in acute infections caused by highly virulent microorganisms. However, we acknowledge that distinguishing true purulence from merely turbid synovial fluid can be challenging. While frank pus is generally unmistakable, cloudy or opaque fluid may also occur in several aseptic conditions, including ALTR, trunnionosis, and crystal or inflammatory arthropathy (McNally et al., 2021; Tande and Patel, 2014; Trebse and Roskar, 2021; Archibeck et al., 2001). Some reports describe ALTR-associated joint effusions as thick, cloudy, “milk-stained”, or yellow-green in appearance (Mikhael et al., 2009; Browne et al., 2010), and similar findings have been observed with other bearing materials (Bonnaig et al., 2011). Kim et al. (2012) also reported yellowish opaque effusion in osteonecrosis of the femoral head. Thus, the presence of turbid fluid around prosthetic components is a characteristic but not a pathognomonic sign of infection. Therefore, in situations where uncertainty exists (such as ALTR or crystal or inflammatory arthropathies), it may, if considered at all, be more appropriately regarded as a suggestive rather than a confirmatory criterion.

In our cohort, the pathogens most frequently associated with purulence were Streptococcus spp. (73 %), Staphylococcus aureus (72 %), Enterobacteriaceae (50 %), and polymicrobial infections (43 %). Only Alijanipour et al. (2015) have previously examined microbial patterns in PJI cases presenting with purulence. They reported Streptococcus spp. (88 %) and Staphylococcus aureus (85 %) as the predominant pathogens associated with pus, which is consistent with our findings.

To our knowledge, this is the first study to report infection-free survival in relation to the presence of purulence. This clinical sign demonstrated a trend towards lower survival rates, although statistical significance was not reached (3-year follow up: 72 % with purulence vs. 82 % without, p=0.068). These findings highlight the need for meticulous deep tissue sampling, surgical debridement, and irrigation in cases presenting with purulence in order to maximize the likelihood of infection eradication and to improve outcomes.

This study is not without limitations. Due to its retrospective character, some variables were missing in certain cases. In particular, the presence of a sinus tract or purulence may be underreported in outpatient, admission, and operation notes. Standardized documentation of intraoperative findings such as sinus tract formation and purulence may improve consistency, reproducibility, and comparability across studies and institutions. The absence of a universally accepted definition of purulence introduces additional variability as its assessment depends on the surgeon's subjective interpretation. Moreover, in cases with a clear preoperative diagnosis of infection, detection bias may influence intraoperative documentation. However, these limitations reflect the everyday challenges of interpreting sinus tracts and purulence and are therefore important to evaluate in this study.

To our knowledge, this is the first study to analyse diagnostic accuracies, microbial profiles, and outcome of clinical criteria in detail, namely sinus tract and purulence surrounding the prosthesis. While there is some literature on sinus tract formation, sufficient evidence on purulence in PJI is almost completely lacking. Larger, prospective studies are necessary to further validate and refine our findings.

While the diagnostic value of a sinus tract has long been recognized and applied in clinical practice for PJI, robust evidence has been lacking. Our study confirms that a sinus tract communicating with the prosthesis exhibits high specificity and can therefore be recommended as a confirmatory criterion for PJI. Similarly, purulence demonstrated excellent specificity, although distinguishing true purulence from turbid synovial fluid can be challenging in clinical practice. Therefore, purulence may, if considered at all, be most appropriately regarded to be a suggestive criterion for PJI.

Sinus tracts were frequently associated with polymicrobial and fungal infections, whereas purulence was most commonly observed in infections caused by high-virulent microorganisms, such as Streptococcus spp. and Staphylococcus aureus. Both sinus tract and purulence were associated with less favourable outcomes, underlining their clinical relevance in the management of PJI.