Leucocyte esterase, glucose and C-reactive protein in the diagnosis of prosthetic joint infections: a prospective study

Open ArchivePublished:March 31, 2016DOI:https://doi.org/10.1016/j.cmi.2016.03.020

      Abstract

      Analysis of joint fluid is of paramount importance for the diagnosis of prosthetic joint infections. Different markers of inflammation and/or infection in joint fluid have been proposed for diagnosis of these infections. In this study we evaluated the performance of leucocyte esterase, C-reactive protein (CRP) and glucose assays in synovial fluids from 129 patients with septic (n = 27) or aseptic (n = 102) prosthetic joint failure. Samples were collected in serum tubes and centrifuged to limit the presence of corpuscle interfering with the assays. Determinations of leucocyte esterase and glucose were carried out by means of enzymatic colorimetric reactions performed on strips for urine analysis. Tests were considered positive when graded + or ++ whereas traces or absence of colour were considered negative. CRP was measured using an automated turbidimetric method and considered suggestive for infections when >10 mg/L. Leucocyte esterase was positive in 25/27 infected patients and negative in 99/102 not infected patients (sensitivity 92.6%, specificity 97.0%). CRP was higher than the threshold in 22/27 infected patients and in 6/102 not infected patients (sensitivity: 81.5%; specificity: 94.1%) whereas glucose showed the lowest sensitivity (77.8%) and specificity (81.4%), being negative in 21/27 and 19/102 infected and not infected patients, respectively. CRP led to a correct diagnosis in 19 of 22 patients with discordant esterase and glucose results. In conclusion, evaluation of leucocyte esterase, glucose and CRP may represent a useful tool for rapid diagnosis of prosthetic joint infections.

      Keywords

      Introduction

      Analysis of joint fluid is of paramount importance for the diagnosis of bone and joint infections, including prosthetic joint infections (PJIs). To date, accurate and efficient diagnosis of bone and joint infections remains one of the most challenging tasks for orthopaedic surgeons and microbiologists. Although several parameters have been proposed for their diagnosis, only a few of them have shown an acceptable accuracy, while others are often time consuming and/or expensive [
      • Tetreault M.W.
      • Wetters N.G.
      • Moric M.
      • Gross C.E.
      • Della Valle C.J.
      Is synovial C-reactive protein a useful marker for periprosthetic joint infection?.
      ,
      • Aggarwal V.K.
      • Tischler E.
      • Ghanem E.
      • Parvizi J.
      Leukocyte esterase from synovial fluid aspirate.
      ,
      • Parvizi J.
      • Adeli B.
      • Zmistowski B.
      • Restrepo C.
      • Greenwald A.S.
      Management of periprosthetic joint infection: the current knowledge: AAOS exhibit selection.
      ,
      • Berbari E.
      • Mabry T.
      • Tsaras G.
      • Spangehl M.
      • Erwin P.J.
      • Murad M.H.
      • et al.
      Inflammatory blood laboratory levels as markers of prosthetic joint infection: a systematic review and meta-analysis.
      ]. Therefore, culture examination is still considered the gold standard, even though it does not appear completely exhaustive, since cultures fail to yield microbial growth in up to 20%–30% of cases [
      • Berbari E.F.
      • Marculescu C.
      • Sia I.
      • Lahr B.D.
      • Hanssen A.D.
      • Steckelberg J.M.
      • et al.
      Culture-negative prosthetic joint infection.
      ].
      To overcome this limitation, determination of markers of inflammation and/or infection directly in joint fluid has been proposed to enhance sensitivity and specificity [
      • Chen A.
      • Fei J.
      • Deirmegian C.
      Diagnosis of periprosthetic infection: novel developments.
      ,
      • Randau T.M.
      • Friedrich M.J.
      • Wimmer M.D.
      • Reichert B.
      • Kuberra D.
      • Stoffel-Wagner B.
      • et al.
      Interleukin-6 in serum and in synovial fluid enhances the differentiation between periprosthetic joint infection and aseptic loosening.
      ]. In particular, leucocyte esterase (LE) has been reported as one of the most promising parameters.
      Leucocyte esterase is an enzyme that catalyses degradation of the connective tissue matrix, which can be found in significant amounts in infected body fluids, typically urine. The secretion of LE by neutrophils recruited to the joint allows for detection of infection using colorimetric reagent strips, such as those used in urine analysis. The test involves a colour change dependent on a hydrolytic reaction that occurs when the enzyme reacts with the chemicals on the test pad. Utility of LE determination for diagnosis of PJI has been confirmed in some studies [
      • Aggarwal V.K.
      • Tischler E.
      • Ghanem E.
      • Parvizi J.
      Leukocyte esterase from synovial fluid aspirate.
      ,
      • Parvizi J.
      • Jacovides C.
      • Antoci V.
      • Ghanem E.
      Diagnosis of periprosthetic joint infection: the utility of a simple yet unappreciated enzyme.
      ,
      • Coiffier G.
      • Pollet S.
      • Albert J.D.
      • Perdriger A.
      • Guggenbuhl P.
      • Chales G.
      Usefulness and limitations of rapid urine dipstick testing for joint-fluid analysis. Prospective single-center study of 98 specimens.
      ,
      • Colvin O.C.
      • Kransdorf M.J.
      • Roberts C.C.
      • Chivers F.S.
      • Lorans R.
      • Beauchamp C.P.
      • et al.
      Leukocyte esterase analysis in the diagnosis of joint infection: can we make a diagnosis using a simple urine dipstick?.
      ,
      • Tischler E.H.
      • Cavanaugh P.K.
      • Parvizi J.
      Leukocyte esterase strip test: matched for musculoskeletal infection society criteria.
      ], which have shown a notable sensitivity and specificity. Moreover, positive LE has been recently proposed as an alternative to synovial leucocyte count among criteria for diagnosis of PJIs [
      • Zmistowski B.
      • Della Valle C.
      • Bauer T.W.
      • Malizos K.N.
      • Alavi A.
      • Bedair H.
      • et al.
      Diagnosis of periprosthetic joint infection.
      ].
      Although showing promising results in some studies, C-reactive protein (CRP) in synovial fluid has received less attention than LE. Nonetheless, it has been suggested as a more accurate method for diagnosing PJIs, compared with its concentration in serum, which may be affected by other inflammatory processes not involving bones or joints [
      • Tetreault M.W.
      • Wetters N.G.
      • Moric M.
      • Gross C.E.
      • Della Valle C.J.
      Is synovial C-reactive protein a useful marker for periprosthetic joint infection?.
      ,
      • Omar M.
      • Ettinger M.
      • Reichling M.
      • Petri M.
      • Guenther D.
      • Gehrke T.
      • et al.
      Synovial C-reactive protein as a marker for chronic periprosthetic infection in total hip arthroplasty.
      ,
      • Vanderstappen C.
      • Verhoeven N.
      • Stuyck J.
      • Bellemans J.
      Intra-articular versus serum C-reactive protein analysis in suspected periprosthetic knee joint infection.
      ].
      More recently, Omar et al. have proposed that combined determination of LE and glucose could improve specificity in diagnosis of septic arthritis [
      • Omar M.
      • Ettinger M.
      • Reichling M.
      • Petri M.
      • Lichtinghagen R.
      • Guenther D.
      • et al.
      Preliminary results of a new test for rapid diagnosis of septic arthritis with use of leukocyte esterase and glucose reagent strips.
      ].
      Therefore, the aim of the present study was to evaluate the applicability of measurement of LE, glucose and CRP in synovial fluids for diagnosis of PJIs in patients with prosthetic joint failure.

      Materials and Methods

       Patient recruitment

      One hundred and twenty-nine consecutive patients who underwent aspiration of synovial fluids for prosthetic joint failure and were referred to Galeazzi Orthopaedic Institute of Milan from February 2014 to May 2015 were studied. Clinically significant PJI was defined as: presence of a sinus tract communicating with the prosthesis or isolation of the same pathogen by culture from at least two separate tissue or fluid samples obtained from the affected prosthetic joints or, finally, having three of the following minor criteria: elevated erythrocyte sedimentation rate and serum CRP concentration, elevated synovial fluid count or positive LE, high percentage of synovial fluid polymorphonuclear neutrophils, positive histology of periprosthetic tissue and single positive culture [
      • Zmistowski B.
      • Della Valle C.
      • Bauer T.W.
      • Malizos K.N.
      • Alavi A.
      • Bedair H.
      • et al.
      Diagnosis of periprosthetic joint infection.
      ]. Patients were included in the study if they had a suspected implant failure of the hip or the knee, a sufficient volume of synovial fluid was obtained and if informed consent was signed. Subjects with other kinds of prostheses (i.e. shoulder) were not considered. Samples were collected as part of standard care. All patients gave their informed consent to participate in the study, which was approved by the scientific committee of the Institute.

       Synovial fluid collection

      Synovial fluids were analysed for CRP, glucose, LE and culture. Samples were collected under aseptic conditions in sterile tubes (for cultures) and in serum tubes containing an acrylic gel and micronized silica (BD Vacutainer SST™; Becton Dickinson, Milan, Italy). Immediately after collection, samples were transported to the laboratory, where they were processed within 5–10 minutes.

       Determination of leucocyte esterase and glucose in synovial fluid

      Determination of LE and glucose in synovial fluid was carried out by means of enzymatic colorimetric strips (Dirui Industrial Co. Ltd, Changchun, China). To limit the presence of corpuscles that could interfere with the assays, samples were centrifuged at 2100 g for 10 min. Then one drop of the supernatant was placed on both LE and glucose pads of one strip; reactions for LE and glucose were read after about 120 and 30 s, respectively, by three of the authors (EDV, FV and MT), who were blinded to the patient's details. Production of purple colour indicated the presence of LE, whereas development of a green colour on the specific pad revealed the presence of glucose. The strip tests were considered positive when graded + or ++ and traces or absence of colour were considered negative. A positive test for LE and one negative for glucose were considered suggestive for infection.

       Determination of C-reactive protein in the synovial fluid

      After sample centrifugation, the synovial CRP was measured by an automated turbidimetric method using a specific reagent kit reacting goat anti-CRP antibody and patient CRP antigen in the sample into sedimentation complexes running on ARCHITECT c4000 system (Abbott Laboratories, Rome, Italy).

       Microbiological cultures

      Synovial fluid was centrifuged at 2100 g for 10 min at room temperature. Prosthetic components and periprosthetic tissues were treated with 0.1% dithiothreitol as previously described [
      • Drago L.
      • Signori V.
      • De Vecchi E.
      • Vassena C.
      • Palazzi E.
      • Cappelletti L.
      • et al.
      Use of dithiothreitol to improve the diagnosis of prosthetic joint infections.
      ]. Briefly, after addition of 0.1% solution of dithiothreitol, and agitation for 15 min, dithiothreitol eluate was centrifuged at 2100 g for 10 min. The supernatant was then discharged and 100 μL of the resuspended pellet were plated on chocolate agar, mannitol salt agar, McConkey agar, Sabouraud agar, and inoculated into brain–heart infusion broth and thioglycollate broth. Plates were incubated for 24–48 h, and broths were maintained for 15 days at 37°C and checked daily for bacterial growth [
      • Drago L.
      • De Vecchi E.
      • Cappelletti L.
      • Vassena C.
      • Toscano M.
      • Bortolin M.
      • et al.
      Prolonging culture to 15 days improves bacterial detection in bone and joint infections.
      ]. When broths became turbid or at the end of the incubation period, aliquots of each broth were plated on chocolate agar and Schaedler (only thioglycollate broth) agar plates.
      Microbial identification was performed on automated instrumentation (Vitek 2 compact; bioMérieux, Marcy l’Etoile, France).

       Statistical analysis

      Differences between infected and not infected patients were evaluated by means of Student's t test for continuous variables (age and CRP) and chi-squared test for categorized data (sex, LE and glucose). For each test, sensitivity, specificity, predictive values and 95% CIs were calculated. Correlation coefficients were also calculated to evaluate the relation between blood and synovial concentrations of CRP and glucose. Strength of synovial CRP measurement was evaluated by calculating the area under the curve (AUC) of the receiver operating characteristics (ROC) curve. Logistic regression analysis was used to identify the accuracy of combinations of variables (LE plus CRP and glucose; LE plus CRP; and LE plus glucose). Differences were considered statistically significant when the p value was ≤0.05. All analyses were carried out on a free-available tool for statistical computation [

      VassarStats: Website for Statistical Computation. Available at www.vassarstats.net. Last accessed December 21, 2015.

      ] and on MedCalc software (MedCalc Statistical Software version 16.2.1; MedCalc Software bvba; Ostend, Belgium; https://www.medcalc.org; 2016).

      Results

      Characteristics of patients are reported in Table 1: the two groups did not significantly differ for male: female ratio or age. Instead, serum CRP and erythrocyte sedimentation rate were markedly higher in septic patients than in aseptic ones. Infected patients had a delayed or late infection; four of them presented an acute onset, whereas in the other patients a chronic PJI was diagnosed.
      Table 1Characteristics of the studied population
      Not infectedInfected
      Number of patients10227
      Age (years) mean ± SD (range)63.8 ± 14.8

      (17–88)
      68.0 ± 12.6

      (38–88)
      Sex (male/female)49/5317/10
      Hip/knee prosthesis33/6912/15
      Serum CRP (mg/L)5.09 (3.49–6.70)49.3 (32.5–66.0)*
      ESR (mm/hr)14.2 (11.8–16.5)41.4 (33.6–49.2)*
      Data of serum C-reactive protein (CRP) and erythrocyte sedimentation rate (ESR) are expressed as mean with 95% CI in parentheses.
      *p <0.001 versus not infected patients.
      Microbiological findings are shown in Table 2, microbial growth was observed in 24 out of 27 infected patients, staphylococci were the most frequently isolated microorganisms followed by streptococci and Enterobacteriaceae.
      Table 2Microorganisms isolated from cultures
      MicroorganismsHip (no. of patients)Knee (no. of patients)
      Staphylococcus aureus14
      Staphylococcus epidermidis23
      Other coagulase negative staphylococci43
      Enterococcus faecalis11
      Streptococcus parasanguinis01
      Streptococcus gordonii10
      Klebsiella pneumoniae20
      Proteus mirabilis10
      Results for LE, glucose and CRP are reported in Table 3. Among not infected patients (n = 102), only three were positive for LE (graded as 1+) whereas in the rest LE was either negative (n = 83) or present in traces (n = 16). By contrast in infected patients (n = 27) LE was scored as negative or present in traces in two patients, classified as 1+ in 12 patients and 2+ in 13 patients (p <0.001, septic versus not septic patients). In the infected group, glucose was negative in 15 patients, graded as traces, 1+, 2+ in six, five and one patient, respectively, while in the not infected group glucose test was negative in ten patients, present in traces in nine patients and scored as 1+ in 69 and as 2+ in 14 patients (p <0.001, infected versus not infected patients). Finally, CRP was significantly higher in infected patients than in patients with no infection (p <0.001). Synovial CRP was positively correlated with serum levels (r = 0.843; p <0.001), but synovial glucose seemed to be unrelated to blood concentrations.
      Table 3Results of leucocyte esterase, C-reactive protein and glucose determination
      Not infectedInfected
      Leucocyte esterase (positive/negative)3/9925/2 *
      Glucose (positive/negative)83/196/21 *
      C-reactive protein (mg/L)3.62 (1.58–5.67)
      a Data are expressed as mean with 95% CI in parentheses.
      37.7 (23.4–52.1)*
      *p <0.001 versus not infected patients.
      a Data are expressed as mean with 95% CI in parentheses.
      The LE test showed the best performances in terms of sensitivity and specificity followed by CRP and glucose as shown in Table 4. The AUC of the ROC curve for CRP was 0.958, suggesting a high accuracy for diagnosis of PJIs (Fig. 1). Logistic regression analysis applied to a model including LE, CRP and glucose showed an AUC of 0.990, which was equal to that of the model combining LE and CRP and higher than the AUC of the ROC curves obtained associating LE and glucose (0.973) or CRP and glucose (0.942).
      Table 4Sensitivity and specificity of leucocyte esterase and glucose test
      Leucocyte esteraseGlucoseC-reactive protein
      Sensitivity (%)92.6 (74.2–98.7)77.8 (57.3–90.6)81.5 (61.2–92.9)
      Specificity97.0 (91.0–99.2)81.4 (72.2–88.1)94.1 (87.1–97.6)
      Positive predictive value89.3 (70.6–97.2)52.5 (36.3–68.1)78.6 (58.5–91.0)
      Negative predictive value98.0 (92.3–99.7)47.5 (32.8–63.7)95.0 (88.3–98.2)
      In parentheses, 95% confidence intervals are reported; for C-reactive protein, the sensitivity and specificity were calculated for a threshold of 10 mg/L.
      Figure thumbnail gr1
      Fig. 1Receiver operating characteristics curve for synovial C-reactive protein.

      Discussion

      Diagnosis of PJIs represents an important challenge for both orthopaedic specialists and microbiologists due to variability in presentation. In fact, a reference standard has not been identified, although guidelines agree on the need to combine laboratory and clinical parameters to make a diagnosis of PJI [
      • Zmistowski B.
      • Della Valle C.
      • Bauer T.W.
      • Malizos K.N.
      • Alavi A.
      • Bedair H.
      • et al.
      Diagnosis of periprosthetic joint infection.
      ,
      • Osmon D.R.
      • Berbari E.F.
      • Berendt A.R.
      • Lew D.
      • Zimmerli W.
      • Steckelberg J.M.
      • et al.
      Infectious Diseases Society of America. Executive summary: diagnosis and management of prosthetic joint infection: clinical practice guidelines by the Infectious Diseases Society of America.
      ].
      Traditionally, counts of synovial leucocytes and differentials have been used in the diagnosis of prosthetic failure and have shown good sensitivity and specificity. However, it requires some time to perform the analysis, and analyses should be performed as soon as possible to avoid the decrease in the leucocyte count that may occur within a few hours [
      • Kerolus G.
      • Clayburne G.
      • Schumacher Jr., J.R.
      Is mandatory to examine synovial fluids promptly after arthrocentesis?.
      ]. Therefore, in the last decade several markers have been evaluated for diagnosis of PJIs directly on synovial fluids, in order to avoid aspecificity of serum parameters, which may reflect conditions that are not strictly related to joint disease. Among these, synovial LE, CRP and α-defensin have received major attention. α-Defensin was excluded from this study, because its measurement by ELISA may require up to 4–5 h, whereas we would like to assess the possibility of using rapid tests for diagnosis of PJIs. We were able to obtain results within 30 min, including sample centrifugation.
      The most relevant finding of our study is the higher positive and negative predictive values of LE with respect to CRP and glucose. To the best of our knowledge, this is the first time that simultaneous determination of LE, CRP and glucose in synovial fluid has been compared in the same patients. LE determination for PJI diagnosis was initially proposed by Parvizi et al. in 2011, who showed its correlation with the percentage of polymorphonuclear leucocytes and total white blood-cell count in synovial fluids [
      • Parvizi J.
      • Jacovides C.
      • Antoci V.
      • Ghanem E.
      Diagnosis of periprosthetic joint infection: the utility of a simple yet unappreciated enzyme.
      ]. In that study, the authors reported a 100% specificity; we obtained a specificity of 98.0%, which is similar to values reported by other authors. A sensitivity of 92.6% was found in our study for LE, which confirms values reported by others [
      • Colvin O.C.
      • Kransdorf M.J.
      • Roberts C.C.
      • Chivers F.S.
      • Lorans R.
      • Beauchamp C.P.
      • et al.
      Leukocyte esterase analysis in the diagnosis of joint infection: can we make a diagnosis using a simple urine dipstick?.
      ,
      • Wetters N.G.
      • Berend K.R.
      • Lombardi A.V.
      • Morris M.J.
      • Tucker T.L.
      • Della Valle C.J.
      Leukocyte esterase reagent strips for the rapid diagnosis of periprosthetic joint infection.
      ]. Infected patients who tested LE negative were diagnosed with hip prosthetic infections; both of them had high synovial CRP, whereas glucose was classified as 1+ and traces, respectively. Cultures of periprosthetic tissues gave Klebsiella pneumoniae and Staphylococcus aureus, while synovial cultures were negative; all other laboratory parameters (erythrocyte sedimentation rate, CRP and histology) and clinical parameters were suggestive for infection.
      The major limit in determination of LE is the difficulty in reading colour development due to the presence of blood or debris [
      • Aggarwal V.K.
      • Tischler E.
      • Ghanem E.
      • Parvizi J.
      Leukocyte esterase from synovial fluid aspirate.
      ,
      • Tischler E.H.
      • Cavanaugh P.K.
      • Parvizi J.
      Leukocyte esterase strip test: matched for musculoskeletal infection society criteria.
      ]. However, this limit can be overcome by centrifuging samples as previously suggested [
      • Aggarwal V.K.
      • Tischler E.
      • Ghanem E.
      • Parvizi J.
      Leukocyte esterase from synovial fluid aspirate.
      ]; collection of synovial fluid in serum tubes containing a gel separator allows avoidance of any interference.
      Synovial CRP showed a lower sensitivity and specificity with respect to LE, but superior to that observed for glucose. CRP synovial concentrations were positively correlated with serum levels. It has been hypothesized that serum CRP is likely to diffuse into joint fluid thanks to alterations in synovial permeability, as a result of the inflammatory process caused by infection [
      • Parvizi J.
      • McKenzie J.C.
      • Cashman J.P.
      Diagnosis of periprosthetic joint infection using synovial C-reactive protein.
      ]. Our data seemed to support this hypothesis.
      Different cut-off values for CRP have been reported ranging between 2.5 and 9.5 mg/L [
      • Tetreault M.W.
      • Wetters N.G.
      • Moric M.
      • Gross C.E.
      • Della Valle C.J.
      Is synovial C-reactive protein a useful marker for periprosthetic joint infection?.
      ,
      • Omar M.
      • Ettinger M.
      • Reichling M.
      • Petri M.
      • Guenther D.
      • Gehrke T.
      • et al.
      Synovial C-reactive protein as a marker for chronic periprosthetic infection in total hip arthroplasty.
      ,
      • Vanderstappen C.
      • Verhoeven N.
      • Stuyck J.
      • Bellemans J.
      Intra-articular versus serum C-reactive protein analysis in suspected periprosthetic knee joint infection.
      ,
      • Parvizi J.
      • McKenzie J.C.
      • Cashman J.P.
      Diagnosis of periprosthetic joint infection using synovial C-reactive protein.
      ], which were associated with variable sensitivity and specificity. In this study, we used a cut-off limit of 10 mg/L, similar to that used by Parvizi et al. (9.5 mg/L) [
      • Parvizi J.
      • McKenzie J.C.
      • Cashman J.P.
      Diagnosis of periprosthetic joint infection using synovial C-reactive protein.
      ], which allows us to obtain acceptable specificity and sensitivity. This shows the need for any laboratory to establish their own cut-off points for synovial CRP, because, as reported by other authors [
      • Tetreault M.W.
      • Wetters N.G.
      • Moric M.
      • Gross C.E.
      • Della Valle C.J.
      Is synovial C-reactive protein a useful marker for periprosthetic joint infection?.
      ], its determination seems to be affected by the analytical method used. Determination of synovial CRP has been recently reported to improve specificity of the α-defensin ELISA test for diagnosis of PJIs by reversing false-positive α-defensin results [
      • Deirmengian C.
      • Kardos K.
      • Kilmartin P.
      • Cameron A.
      • Schiller K.
      • Parvizi J.
      Combined measurement of synovial fluid α-defensin and C-reactive protein levels: highly accurate for diagnosing periprosthetic joint infection.
      ]. Similarly in the present study, synovial fluids yielding false-positive LE results showed low CRP values, although concordance between the CRP and LE tests was found in only 89.1% of cases. Glucose and LE assays were concordant in 85.1% of patients, whereas concordance between CRP and glucose was observed in an even lower rate of cases (77.5%).
      Recently, combined evaluation of LE and glucose has been proposed for diagnosis of septic arthritis [
      • Omar M.
      • Ettinger M.
      • Reichling M.
      • Petri M.
      • Lichtinghagen R.
      • Guenther D.
      • et al.
      Preliminary results of a new test for rapid diagnosis of septic arthritis with use of leukocyte esterase and glucose reagent strips.
      ]. As they may be assessed on the same reagent strip without added costs, we decided to evaluate the usefulness of glucose determination for diagnosis of PJIs. Our results are in contrast with those of Omar et al.: glucose determination did not improve sensitivity or specificity of the LE test and, taken alone, it showed the lowest sensitivity and specificity. However, we must consider that in respect to the study of Omar et al., we evaluated only patients with suspected PJIs, and not those with septic arthritis of the native joint. In contrast to what was observed with CRP, synovial glucose seemed to be unrelated to blood concentrations. To the best of our knowledge, correlation between synovial and blood glucose has never been deeply investigated. However, patients with elevated blood glucose concentrations did not show a parallel increase in synovial glucose levels.
      According to logistic regression analysis, determination of glucose did not increase the accuracy of combined evaluation of LE and PCR, as shown by the AUC of the ROC curves. Considering the longer turn-around time and higher cost of CRP determination with respect to determination of LE and glucose, it could be hypothesized to use it in cases of discordance between assays for LE and glucose. Applying this rule, in the present study, 19 of 22 patients with discordant LE and glucose tests could have been correctly diagnosed.
      This study has obviously some limitations: the first one is the limited infected population enrolled in the study, which therefore needs further confirmation. It has also to be considered that, as our centre is dedicated to treatment of bone and joint infections, the prevalence of infections in our population might be higher than that observed in other hospitals, so altering the positive predictive value, which is influenced by the prevalence of the disease.
      In conclusion, determination of LE and glucose in synovial fluid together with that of CRP may be a valid support in the diagnosis of PJIs and their measurements could be transferred to the clinical setting.

      Transparency Declaration

      The authors declare non conflicts of interest.

      Contribution to Authorship

      EDV, CLR and LD conceived and designed the study and revised the manuscript; EDV and FV drafted the article; EDV, FV, MB and MT performed the analyses and analysed the data; LT critically revised the manuscript; LD approved the final version to be submitted.

      References

        • Tetreault M.W.
        • Wetters N.G.
        • Moric M.
        • Gross C.E.
        • Della Valle C.J.
        Is synovial C-reactive protein a useful marker for periprosthetic joint infection?.
        Clin Orthop Relat Res. 2014; 472: 3997-4003
        • Aggarwal V.K.
        • Tischler E.
        • Ghanem E.
        • Parvizi J.
        Leukocyte esterase from synovial fluid aspirate.
        J Arthroplasty. 2013; 28: 193-195
        • Parvizi J.
        • Adeli B.
        • Zmistowski B.
        • Restrepo C.
        • Greenwald A.S.
        Management of periprosthetic joint infection: the current knowledge: AAOS exhibit selection.
        J Bone Joint Surg. 2012; 94: e104
        • Berbari E.
        • Mabry T.
        • Tsaras G.
        • Spangehl M.
        • Erwin P.J.
        • Murad M.H.
        • et al.
        Inflammatory blood laboratory levels as markers of prosthetic joint infection: a systematic review and meta-analysis.
        J Bone Joint Surg Am. 2010; 92: 2102-2109
        • Berbari E.F.
        • Marculescu C.
        • Sia I.
        • Lahr B.D.
        • Hanssen A.D.
        • Steckelberg J.M.
        • et al.
        Culture-negative prosthetic joint infection.
        Clin Infect Dis. 2007; 45: 1113-1119
        • Chen A.
        • Fei J.
        • Deirmegian C.
        Diagnosis of periprosthetic infection: novel developments.
        J Knee Surg. 2014; 27: 259-265
        • Randau T.M.
        • Friedrich M.J.
        • Wimmer M.D.
        • Reichert B.
        • Kuberra D.
        • Stoffel-Wagner B.
        • et al.
        Interleukin-6 in serum and in synovial fluid enhances the differentiation between periprosthetic joint infection and aseptic loosening.
        PLoS One. 2014; 9: e89045
        • Parvizi J.
        • Jacovides C.
        • Antoci V.
        • Ghanem E.
        Diagnosis of periprosthetic joint infection: the utility of a simple yet unappreciated enzyme.
        J Bone Joint Surg Am. 2011; 93: 2242-2248
        • Coiffier G.
        • Pollet S.
        • Albert J.D.
        • Perdriger A.
        • Guggenbuhl P.
        • Chales G.
        Usefulness and limitations of rapid urine dipstick testing for joint-fluid analysis. Prospective single-center study of 98 specimens.
        Joint Bone Spine. 2013; 80: 604-607
        • Colvin O.C.
        • Kransdorf M.J.
        • Roberts C.C.
        • Chivers F.S.
        • Lorans R.
        • Beauchamp C.P.
        • et al.
        Leukocyte esterase analysis in the diagnosis of joint infection: can we make a diagnosis using a simple urine dipstick?.
        Skeletal Radiol. 2015; 44: 673-677
        • Tischler E.H.
        • Cavanaugh P.K.
        • Parvizi J.
        Leukocyte esterase strip test: matched for musculoskeletal infection society criteria.
        J Bone Joint Surg Am. 2014; 96: 1917-1920
        • Zmistowski B.
        • Della Valle C.
        • Bauer T.W.
        • Malizos K.N.
        • Alavi A.
        • Bedair H.
        • et al.
        Diagnosis of periprosthetic joint infection.
        J Arthroplasty. 2014; 29: 77-83
        • Omar M.
        • Ettinger M.
        • Reichling M.
        • Petri M.
        • Guenther D.
        • Gehrke T.
        • et al.
        Synovial C-reactive protein as a marker for chronic periprosthetic infection in total hip arthroplasty.
        Bone Joint J. 2015; 41: 173-176
        • Vanderstappen C.
        • Verhoeven N.
        • Stuyck J.
        • Bellemans J.
        Intra-articular versus serum C-reactive protein analysis in suspected periprosthetic knee joint infection.
        Acta Orthop Belg. 2013; 79: 26-30
        • Omar M.
        • Ettinger M.
        • Reichling M.
        • Petri M.
        • Lichtinghagen R.
        • Guenther D.
        • et al.
        Preliminary results of a new test for rapid diagnosis of septic arthritis with use of leukocyte esterase and glucose reagent strips.
        J Bone Joint Surg Am. 2014; 96: 2032-2037
        • Drago L.
        • Signori V.
        • De Vecchi E.
        • Vassena C.
        • Palazzi E.
        • Cappelletti L.
        • et al.
        Use of dithiothreitol to improve the diagnosis of prosthetic joint infections.
        J Orthop Res. 2013; 31: 1694-1699
        • Drago L.
        • De Vecchi E.
        • Cappelletti L.
        • Vassena C.
        • Toscano M.
        • Bortolin M.
        • et al.
        Prolonging culture to 15 days improves bacterial detection in bone and joint infections.
        Eur J Clin Microbiol Infect Dis. 2015; 34: 1809-1813
      1. VassarStats: Website for Statistical Computation. Available at www.vassarstats.net. Last accessed December 21, 2015.

        • Osmon D.R.
        • Berbari E.F.
        • Berendt A.R.
        • Lew D.
        • Zimmerli W.
        • Steckelberg J.M.
        • et al.
        Infectious Diseases Society of America. Executive summary: diagnosis and management of prosthetic joint infection: clinical practice guidelines by the Infectious Diseases Society of America.
        Clin Infect Dis. 2013; 56: 1-10
        • Kerolus G.
        • Clayburne G.
        • Schumacher Jr., J.R.
        Is mandatory to examine synovial fluids promptly after arthrocentesis?.
        Arthritis Rheum. 1989; 32: 271-278
        • Wetters N.G.
        • Berend K.R.
        • Lombardi A.V.
        • Morris M.J.
        • Tucker T.L.
        • Della Valle C.J.
        Leukocyte esterase reagent strips for the rapid diagnosis of periprosthetic joint infection.
        J Arthroplasty. 2012; 27: 8-11
        • Parvizi J.
        • McKenzie J.C.
        • Cashman J.P.
        Diagnosis of periprosthetic joint infection using synovial C-reactive protein.
        J Arthroplasty. 2012; 27: 12-16
        • Deirmengian C.
        • Kardos K.
        • Kilmartin P.
        • Cameron A.
        • Schiller K.
        • Parvizi J.
        Combined measurement of synovial fluid α-defensin and C-reactive protein levels: highly accurate for diagnosing periprosthetic joint infection.
        J Bone Joint Surg Am. 2014; 96: 1439-1445