Combined Serum and Synovial C-Reactive Protein Tests: An Economical and Ecient Adjunct to the Diagnosis of Chronic Prosthetic Joint Infection

Background Diagnosis of Periprosthetic joint infection (PJI) is very complex and challenging, especially for chronic PJI. The value of C-reactive protein (CRP) in infectious diseases has been recognized, but the diagnostic value of CRP in chronic PJI is unknown. Our objective was to investigate the effectiveness of synovial CRP in chronic PJI and to determine the optimal combination of serum and synovial CRP in distinguishing chronic PJI from aseptic failure after knee and hip arthroplasties. Methods From January 2018 to December 2019, we prospectively included patients scheduled to have a revision surgery for chronic PJI or aseptic loosening of an implant, in which synovial CRP was additionally measured along with routine preoperative diagnostic serum and synovial biomarkers. The receiver operating characteristic (ROC) curves and area under the curve (AUC) were analyzed for each biomarker to determine diagnostic ecacy. Results There were no statistically signicant differences in demographic data among the 97 cases we eventually included. the synovial CRP levels were signicantly higher in the infection group than in the aseptic group (median: 19 mg/l vs. 9.25 mg/l; p = .001). The optimal cut-off value for detecting chronic PJI of synovial CRP was of 7.26 mg/l with a sensitivity of 84.62%, a specicity of 93.10%. The combined model I (Serum CRP > 10.2 mg/l OR SF CRP > 7.2 6 mg/l) had a negative predictive value (NPV) of 96.67%, and a sensitivity of 97.44%. The combined model II (Serum CRP > 10.2 mg/l AND Synovial CRP > 7.26 mg/l) led to a specicity of 1, and a positive predictive value (PPV) of 1. Conclusion The present study demonstrated that the combination of serum and synovial CRP can be used as an adjunct to the diagnosis of chronic PJI.


Introduction
Prosthetic joint infection (PJI) is a devastating complication that can occur following total joint replacement [1]. However, discrimination between infected and aseptic failed total joint replacements can be di cult in some cases, especially in patients with chronic PJI [2]. Establishing a diagnosis of chronic PJI is challenging due to atypical symptoms, which may lead to infection with delayed healing, severe bone defects, joint dysfunction and even a higher risk of short-term mortality [3]. To date, there was no "gold standard" tests or protocol for diagnosing chronic PJI [4]. So, an accurate and timely diagnosis of chronic PJI is a key step toward implementing an effective treatment.
Patients with chronic PJI, elevated serum in ammatory biomarkers, such as serum C-reactive protein (CRP) and erythrocyte sedimentation rate (ESR) may be the primary indications of PJI due to atypical clinical symptoms such as joint effusion, pain, swelling, and redness. However, due to insu cient speci city, these tests are not su cient to diagnose PJI alone [5]. They also need to be supplemented by more speci c tests, such as synovial uid analysis, microbial culture, and histopathology. Among them, the detection of synovial uid (such as IL-6, CD64, and alpha-defensin) has shown great attraction in recent studies [4,6,7]. CRP is one of the most widely used in ammatory markers in the identi cation of infectious diseases, which is synthesized primarily in liver hepatocytes but also by smooth muscle cells, macrophages, endothelial cells, lymphocytes, and adipocytes [8,9]. Its physiological role is to bind to lysophosphatidylcholine expressed on the surface of dead or dying cells (and some types of bacteria) in order to activate the complement system via C1q [10]. C-reactive protein is a marker for in ammation, and its levels increase during bacterial infection. And studies have shown that, CRP is deposited at sites of in ammation and tissue damage in both naturally occurring and experimental conditions [11]. Therefore, the deposition of CRP at the site of infection can be detected, especially in the liquid phase [12]. Recent studies have shown that measuring CRP levels in synovial uid may be a valuable and cost-effective means to improve the diagnosis of PJI [1,13,14]. However, synovial CRP studies to date were limited to small sample sizes, and studies on chronic low-toxicity PJI were insu cient.
In this study, we thus sought to (1) determine the utility of serum and synovial CRP in distinguishing between aseptic failure and chronic infections in patients undergoing revision surgery for failure of total joint arthroplasty, and (2) establish combined cut-off values of serum and synovial CRP in con rming chronic PJI.

Patients And Methods
Before beginning this study, we obtained Institutional Review Board approval for collection of all patient samples. We enrolled patients who were scheduled to have a revision surgery for indications of chronic infection of knee and hip arthroplasties or aseptic loosening of an implant from January 2019 to December 2020. The patients were divided into two groups, "aseptic" and "infection", based on the 2013 Musculoskeletal Infection Society (MSIS) criteria for the diagnosis of PJI [15]. The "aseptic revision" group was de ned as patients who did not ful ll the de nition of PJI and did not develop infection or undergo a reoperation for at least 1 year following the index arthroplasty. A postoperative infection was considered 'chronic' when PJI symptoms occurred beyond 6 weeks after implantation [16][17][18].
To rule out diseases associated with elevated in ammatory markers, patients with the following conditions were excluded: malignancy, rheumatism, renal failure, autoimmune disease, chronic infectious disease (such as human immunode ciency virus or hepatitis C virus), and patients with recent antibiotic use (less than two weeks).
All synovial uid aspirates were collected either preoperatively or in the operating room at the time of revision surgery. And blood samples were obtained preoperatively. To assess CRP levels, plasma and synovial uid were stored in lithium-heparin vacuum collection tubes. The CRP was tested using a particle-enhanced turbidimetric immunoassay with a HITACHI 7600 Series Automatic Biochemical Analyzer (Hitachi, Tokyo, Japan) and diagnostic kit (DiaSys Diagnostic Systems GmbH, Shanghai, China). Synovial uid was examined for WBCs and PMNs using a haematology analyzer (Symex XE-5000 haematology analyzer, Symex, Japan), and the synovial uid is cultured for 14 days on Columbia agar, chocolate agar, and Schaedler agar. At least three suspected tissue specimens were obtained intraoperatively by a stationary surgeon for culture, and the tissue specimens were subjected to intraoperatively frozen section and histopathological examination.

Statistical Analysis
Data were recorded using Microsoft Excel (Microsoft Corporation, Richmond, VA), and statistical analysis was carried out using SPSS version 24. The data are presented as medians and interquartile ranges (IQRs). The results of the diagnostic tests were compared between the groups using an independentsamples t-test. We used receiver operating characteristic (ROC) curve analyses, with Wald CIs, to assess the ability of serum and synovial uid CRP concentration to determine the presence of PJI. Youden's J statistic was used to determine optimum cut-off values for the diagnosis of chronic infection. The area under the curves (AUCs) of each test were compared using MedCalc 13.2.2 Software (MedCalc Software BV, Ostend, Belgium), and based on the cut-off values, the sensitivity, speci city, positive predictive value (PPV), and negative predictive value (NPV) of these makers were calculated from contingency tables. A pvalue ≤ 0.05 was considered statistically signi cant.
The mean concentration of CRP in the synovial uid for the infection group was 9.93 mg/ml and was signi cantly higher (p < 0.001) than the aseptic group with a mean concentration of 3.58 mg/ml ( Table  2). Serum CRP and synovial uid percentage of polymorphonuclear neutrophils (PMN%) were signi cantly higher in the infection group than in the aseptic group (Table 2 and Fig. 1). While the ESR did not differ signi cantly between the two groups (p=.097).
To visualize the sensitivity and speci city of the measured biomarkers to predict the cause of revision (aseptic or infection), a conventional ROC curve and the area under the curve (AUC) were calculated ( Figure 2). The ROC curve analysis revealed the highest AUC for synovial uid CRP, at 0.937 (95% con dence interval (95% CI), 0.869-0.976). Using Youden's index, the optimal cut-off values were 7.26 mg/l, 10.2 mg/l, 69.79% and 34 mm/h, for synovial CRP, serum CRP, synovial PMN% and ESR, respectively, discriminating between PJI and aseptic failure. Table 3, the synovial CRP level (7.26 mg/l) demonstrated a mean sensitivity of 84.62% (95% CI 69.5 to 94.1) and a mean speci city of 93.10% (95% CI 83.3% to 98.1%). The optimal serum CRP cutoff value was calculated at 10.02 mg/l, with sensitivity, speci city, and negative predictive value (NPV) of  (Table 3).

Discussion
In this prospective study, we analyzed and compared serum and synovial in ammatory factors in patients with chronic infection and aseptic after knee and hip arthroplasties. Our data indicate that serum and synovial CRP was signi cantly higher in the chronic infected group than in the aseptic group (35 mg/l vs 21.5 mg/l, and 9.93 mg/l vs 3.58 mg/l, p < 0.001, Table 2). We found a strong correlation between serum and synovial uid CRP level (r = 0.523, P = .0012). This suggests that the circulatory system CRP is detected by spreading to the synovial uid through increased vascular and synovial permeability due to infection [13,19]. We determined that the serum CRP threshold for diagnosing chronic PJI was 10mg/l, which was signi cantly lower than the results (39.8 mg/l) of a recent multicentric study conducted by Parvizi [20]. We found that when the threshold of synovial CRP was 7.26 mg/l, the AUC area of chronic PJI was as high as 93.70% (95 CI 0.869 to 0.976). However, the thresholds for synovial CRP that we determined were different from previous studies [13,21] (2.8mg/l to 9.5mg/ml), which may be due to reduced in ammatory response in chronic PJI patients with low-toxic infections, as well as differences in measurement.
Meanwhile, we also show that when serum CRP > 10.2mg/l or synovial CRP > 7.26mg/l, the diagnosed NPV value reached 96.67%. This combination can signi cantly improve diagnostic sensitivity, but speci city will be sacri ced. In our cohort, there will be 29 aseptic patients who will be misdiagnosed with an infection. Accordingly, when serum and synovial CRP were both high than 10.2 mg/l and 7.26 mg/l, respectively, the diagnostic PPV value reached 1, and the speci city of diagnosis was improved to 1. This combination can be prepared to exclude non-infected patients and improve the e ciency of clinical diagnosis. To our knowledge, this is the rst study to use different predictive models of serum and synovial CRP in the identi cation of chronic PJI.
The sensitivity of serum CRP reached 84.62% was in line with that of a previously published study, included 4934 participants, by Parvizi J et al [22]. However, the serum CRP showed false-positive cases and, hence, a reduced speci city in our study, resulting in a potential overtreatment with unnecessary surgical revisions and prolonged antimicrobial treatment if used alone. The reasonable explanation is that CRP, a factor secreted by a variety of cells, increases to different degrees when the body is stimulated by in ammation [3,5]. Therefore, we investigated in ammatory markers of local synovial uid in joints with higher speci city [4]. Based on previous studies, detection time and economic bene ts, we focused on the value of synovial CRP in differentiating chronic PJI [13,14,23,24]. Our study indicated the synovial CRP possess a 93.10% diagnostic speci city and 89.69% diagnostic accuracy in identifying chronic PJI (Table 3). However, the analysis found that synovial-CRP sensitivity in differentiating chronic PJI was only 84.62%, ranging from 69.5% to 94.1%. A possible reason for the low sensitivity is the formation of mature bio lms in patients with chronic infection, which protect the pathogen against the host immune system resulting in a weakened immune response and, hence, reduced release of CRP [25,26]. Therefore, we also do not recommend the use of synovial CRP alone for the diagnosis of chronic PJI. Therefore, we recommend the combined application of serum CRP and synovial CRP to timely detect chronic PJI. The current data suggested that the combination model I of serum and synovial CRP (Serum CRP > 10.2 OR Synovial CRP > 7.26) be rst used to achieve higher diagnostic sensitivity and to reduce the false negative rate of diagnosis. And then using the combined model II of serum and synovial CRP (Serum CRP > 10.2 AND Synovial CRP > 7.26) to improve diagnostic speci city and thus reduce the false positive rate. The effectiveness of this joint diagnostic approach was recently recognized in a systematic review by Abdelbary et al [27]. After this screening, histopathology and other tests can be used in the remaining patients to minimize missed and misdiagnosis. There was no consensus on the optimal thresholds of serum CRP alone and its combination with synovial CRP for the detection of chronic PJI.
The current study established thresholds for serum and synovial CRP (10.2mg/l and 7.26mg/l, respectively), and developed two predictive models for the diagnosis of chronic PJI that were highly valuable. In addition, based on the premise of effectiveness, serum and synovial CRP have low costeffectiveness, (USD $12 per test) and the detection ability of CRP is available in many hospitals. Therefore, this combination method has the potential to be widely used.
There were some limitations in the present study. First of all, there is no clear consensus on the exact time of postoperative chronic PJI. Although the time of bio lm maturation and previous published studies were used as references in this study, cases of acute infection in included chronic PJI patients could not be completely excluded [4,28,29]. Second, this study involved a single center, and the sample size was relatively small, with only 39 cases in the chronic PJI group and 58 cases in the aseptic group. However, this preliminary trial shows valuable results and warrants a larger multicentric study to verify the e cacy of serum and synovial CRP in the diagnosis of chronic PJI. Finally, patients with recent antibiotic use were excluded from this study for the elimination of confounding factors. However, this may reduce the number of patients enrolled, limiting the generalizability of the results of this study.

Conclusions
This is the rst study concentrating on the diagnoses of even low-grade infections against aseptic failure of knee and hip arthroplasty. Synovial CRP is a valuable test that can well exclude chronic PJI through a combination of serum and synovial CRP (Serum CRP ≤ 10.2mg/l AND Synovial CRP ≤ 7.26mg/l). Therefore, this study is a valuable addition to the current diagnostic criteria developed by the International Consensus Conference on Musculoskeletal Infection (ICM) [30]. And we concluded that preoperative testing for ESR had no bene t for the diagnosis of chronic PJI.

Declarations Author contributions
Wei Huang: Conceptualized the study, Carried out the statistical analysis, Reviewed the manuscript, Lead author of original trial data.
Hai Wang: Collected and analyzed the data, Wrote the manuscript.
Leilei Qin and Ning Hu: Carried out the statistical analysis, Reviewed the manuscript.
Jiawei Wang: Collected and analyzed the data.

Funding
Not applicable.

Availability of data and materials
We do not wish to share our data, because some of the patient's data regarding individual privacy, and according to the policy of our hospital, the data could not be shared with others without permission. An anonymised form of the data could be made available from the corresponding author upon reasonable request.   Cells   Table III. Sensitivity, Speci city, PPV, NPV, and accuracy of in ammatory markers