The Journal of Arthroplasty 29 (2014) 1105–1109
Contents lists available at ScienceDirect
The Journal of Arthroplasty journal homepage: www.arthroplastyjournal.org
Synovial IL-6 AS Inflammatory Marker in Periprosthetic Joint Infections Markus Lenski, Cand. med. a, b, Michael A. Scherer, Dr. med. a a Department of Orthopedics and Trauma Surgery, Klinikum Dachau, Academic Teaching Hospital of the Ludwig-Maximilians-University of Munich, 7 Krankenhausstraße 15, 85221 Dachau, Germany b Faculty of Medicine, Technical University of Munich, Munich, Germany
a r t i c l e
i n f o
Article history: Received 17 June 2013 Accepted 14 January 2014 Keywords: periprosthetic joint infection inflammatory marker interleukin-6 synovial fluid synovial fluid white blood cell count lactate
a b s t r a c t We analyzed serum and synovial biomarkers of 69 patients. 31 of them suffered from a periprosthetic joint infection (PJI) and 38 from aseptic arthralgia after total joint arthroplasty. We used Receiver-OperatingCharacteristic-curves to calculate the Area-under-the-curve (AUC), cutoff-values, positive (+LR), negative (− LR) and interval-Likelihood-Ratios (iLR) for predicting a PJI. The most significant parameter was synovial interleukin-6 (IL-6) (cutoff-value ≥ 30,750 pg/ml, AUC = 0.959, SE = 90.0%, SP = 94.7%, + LR = 17.27), followed by synovial lactate (cutoff-value ≥ 8.3 mmol/l, AUC = 0.844, SE = 71.4%, SP = 88.0%, + LR = 5.95), and synovial glucose (cutoff-value ≤ 44 mg/dl, AUC = 0.829, SE = 79.2%, SP = 78.6%, +LR = 3.69). IL-6 ≥30,750 pg/ml and lactate ≥ 10 mmol/l make a PJI very likely, IL-6 b10,000 pg/ml or lactate b 4.3 mmol/l makes a PJI very unlikely. If none of these thresholds are met, physicians should use the iLR of IL-6, glucose and lactate to estimate the likelihood of PJI. © 2014 Elsevier Inc. All rights reserved.
The demographic change in Europe and North America as well as the medical progress in the field of arthroplasty leads to an increasing number of total joint arthroplasties, and with them the number of periprosthetic joint infections (PJIs) increases [1]. After total joint arthroplasty, an infection of the prosthesis occurs in 0.7% [2] and is a serious complication leading to revision arthroplasty, multiple surgical interventions, long-lasting disability of the patient and poor functional outcome [3,4]. Revision total joint arthroplasties are responsible for longer mean length of hospitalization, higher complication rates, for elevated mortality, especially in elderly patients and for the concomitant effect of increased hospital costs [3,5]. On account of this the diagnosis of PJI has to be set up rapidly and reliably, thus today’s research in the field of PJI concentrates on biomarkers that allow diagnosing PJIs with high sensitivity, specificity and accuracy. Deirmengian et al [6] and Jacovides et al [7] evaluated the role of various biomarkers in the synovial fluid in differing PJI from aseptic failure of prosthesis. The currently used biomarkers in daily clinical routine are either highly sensitive and less specific (e.g. polymerase chain reaction [8,9], erythrocyte sedimentation rate, C-reactive Protein [CRP] and peripheral white blood cells [pWBC] [10–12]) or highly specific and less sensitive (e.g. gram stain [13–15], synovial fluid white blood cell count [SFWBC] [16–18], synovial glucose[18]) and some show low diagnostic accuracy (e.g. synovial LDH, synovial total protein [18]). Deirmengian et al [6] and Jacovides et al [7] The Conflict of Interest statement associated with this article can be found at http://dx.doi.org/10.1016/j.arth.2014.01.014. Reprint requests: Markus Lenski, Gerhard Hanke Weg 22, 85221 Dachau, Germany. http://dx.doi.org/10.1016/j.arth.2014.01.014 0883-5403/© 2014 Elsevier Inc. All rights reserved.
revealed that especially the inflammatory marker IL-6 seems to have high diagnostic potential and accuracy for diagnosing PJIs. The aim of this study was (1) to investigate the role of synovial IL-6 in diagnosing PJI in daily clinical routine and (2) to compare its diagnostic potential to standard inflammatory parameters in the serum and in the synovial fluid in suspected PJI. The hypothesis was that IL-6 would be more useful for differing between both pathologic conditions than standard inflammatory markers and would be the first inflammatory parameter with excellent diagnostic potential for predicting PJI, when compared to patients with aseptic arthralgia after total joint arthroplasty (TJA). Patients and Methods 69 Patients that had undergone primary total joint arthroplasty and presented to the emergency unit of our hospital with symptoms consistent with a PJI, such as pain, swelling, redness, warming, impaired range of motion and fever, were included in this study. On admission, blood samples were taken to define the concentration of the C-reactive Protein (CRP) and the peripheral white blood cell count (pWBC) in a routine laboratory. The suspected joint was aspirated under strictly aseptic conditions and the levels of lactate, glucose, LDH and the synovial fluid white blood cell count (SFWBC) were examined in a clinical routine laboratory. In addition, the levels of interleukin-6 (IL-6) in the synovial fluid were determined by using the IMMUNOLITE 1000® Immunoassay System (SIEMENS Healthcare, Germany). The commercially available assay was optimized for synovial fluid, but not especially altered to account for synovial fluid matrix effects. Several dilution steps were necessary in order to account for the extremely elevated synovial levels compared to
1106
M. Lenski, M.A. Scherer / The Journal of Arthroplasty 29 (2014) 1105–1109
serum levels of IL-6. All assays were carried out according to the manufacturer’s instructions. The diagnosis criteria used for PJI were modeled after the Musculoskeletal Infection Society (MSIS) [19]. These criteria included a communicating sinus tract with the prosthesis, or a microbiologic test of the synovial fluid showing a positive microbiological culture, or at least four of the following five criteria: presence of pus in the affected joint, histological proof of infection, CRP level N 10 mg/dl [20], pWBC N 11 × 10 3/μl [21] and SFWBC N 20 × 10 3/μl [19,20,22] (n = 31). The control group (n = 38) consisted of patients who complained about aseptic arthralgia after TJA and had a CRP of less than 10 mg/dl, an SFWBC of less than 20 × 10 3/μl [19,22] and at least one negative microbiological culture. Arthralgia is the leading symptom in PJIs. In patients with arthralgia after TJA an infection was ruled out during their stay in hospital. The Receiver-Operating-Characteristic-curve (ROC) and the corresponding Area-under-the-curve (AUC) were calculated for each inflammatory parameter. Parameters with an AUC N 0.9 have an excellent and those with an AUC N 0.8 a good diagnostic potential. Optimal cutoff-values were determined by maximizing sensitivity and specificity (Youden’s J statistic). Cutoff-values lead to false positive and false negative test-result, leading to a loss of information of the obtained data, which is problematic for diagnosing PJIs. This is the reason why Likelihood-Ratios (LRs) and interval Likelihood ratios (iLRs) were calculated. Furthermore LRs and iLRs are more useful than sensitivity and specificity for clinical bedside application [23]. LRs are used to estimate the change from pretest to posttest probability of a disease. An LR above 1 increases the posttest-probability of a disease and an LR lower 1 decreases it. To be more specific, an LR of 5 increases the posttest disease probability + 30% and an LR of 10 increases it + 45%, whereas an LR of 0.5 decreases the posttest probability − 15%, an LR of 0.1 decreases it − 45% [24]. These estimated changes of the pretest to posttest probability can only be used, if the pretest probability of a disease is between 10% and 90%. The positive Likelihood-Ratio (+ LR) gives the change from pretest to posttest probability for a positive test result, the negative LikelihoodRatio (− LR) gives the change for a negative test result. Interval Likelihood-Ratios are given on a continuous scale of biomarker levels and use more of the contained information of inflammatory markers than LRs and therefore they are more useful for the estimation of the disease probability as well as for clinical decision making. The whole statistical analysis was performed by using SPSS for Windows (SPSS 17.0, Inc., Chicago, IL). A statistics institution affirmed our statistical approach. Participants of this clinical trial presented to our emergency unit of the hospital from October 2009 until October 2012. This study was conducted in accordance to the Declaration of Helsinki. No approval of the Ethic Review Committee of the Technical University of Munich, Germany, was sought for this retrospective epidemiologic study without personal data, as all data were only obtained if medically indicated. No blood test or arthrocentesis was done for the sole purpose of this study. Diagnostic procedures and therapeutic treatment were not directly influenced. The established standard operating workflow was used in case of suspicion of septic arthritis, providing a solid basis for routine data collection. Results 69 Patients took part in this clinical trial, 31 of them suffered from a perisprosthetic joint infection. The isolated pathogens in the PJI were Staphylococcus aureus (n = 11), coagulase negative staphylococci (n = 4), β-hemolytic streptococci (n = 5), gram negative bacteria (n = 8), Finegoldia magna (n = 4) and Enterococcus faecalis (n = 1). In two cases of PJIs, the growth of two pathogens was observed. 20 knee prostheses, 10 hip prostheses and 1 elbow prosthesis were infected. Arthralgia after TJA was the diagnosis in 38 patients. The
most frequently affected joint was the knee (n = 27), followed by the hip (n = 9) and the elbow (n = 2). The mean concentration of IL-6 in the synovial fluid during a PJI was 308,099 pg/ml and was statistically significantly higher (P = 0.01) than during arthralgia after TJA with a mean concentration of 10,333 pg/ml. The AUC was 0.959 (Fig. 1) with an optimal threshold at 30,750 pg/ml (sensitivity = 90.9%, specificity = 94.7%, + LR = 17.27, − LR = 0.10) (Table 1). The calculation of the interval Likelihood-Ratio showed that IL-6 levels below 10,000 pg/ml drastically decreased the posttest-probability of PJI (iLR = 0.08), IL-6 levels above 49,000 pg/ml led to a substantial change of the posttest probability of PJI (iLR = 13.82) (Table 2). The mean glucose level in the synovial fluid of patients with PJI was 30.4 mg/dl and statistically significantly (P = 0.001) decreased compared to patients with arthralgia after TJA with a mean of 71.2 mg/dl. The AUC of synovial glucose was 0.829, revealing an optimal cutoff-value at 44 mg/dl (sensitivity = 79.2%, specificity = 78.6%, + LR = 3.69, − LR = 0.27). The iLR revealed that glucose levels above 90 mg/dl decrease the likelihood of PJI, whereas glucose levels below 30 mg/dl increased it (Table 2). The arithmetical mean of lactate levels in the synovial fluid during PJI was statistically significantly (P = 0.001) increased compared to levels during arthralgia after TJA with a mean of 12.8 mmol/l compared to 5.7 mmol/l. The determined AUC of lactate was 0.844, leading to an optimal threshold of 8.3 mmol/l (sensitivity = 71.4%, specificity = 88.0%, + LR = 5.95, − LR = 0.33). The posttest probability of PJI decreased considerably at lactate levels below 4.3 mmol/l (iLR = 0) and increased drastically at lactate levels of 10 mmol/l or higher (iLR = 14.29). LDH levels in the synovial fluid had an AUC of 0.805 and showed a mean of 4918 U/l during PJI, which was statistically significantly (P = 0.007) increased compared to the mean of 1832 U/l during arthralgia after TJA. The best cutoff-value was 1423 U/l (sensitivity = 84.0%, specificity = 72.0%, + LR = 3.00, − LR = 0.22). The AUC of the SFWBC was 0.807, the mean SFWBC was 69.6 × 10 3/μl during PJI and 6.3 × 10 3/μl during arthralgia after TJA. The arithmetical means of both groups differed statistically significantly (P = 0.000008). The best threshold was an SFWBC of 23.0 × 10 3/μl (sensitivity = 60.0%, specificity = 94.3%, + LR = 10.50, − LR = 0.42). By using a threshold of 50 × 10 3/μl, the sensitivity was 52% and the specificity became 100%. At an SFWBC below 10 × 10 3/μl the likelihood of PJI decreased (iLR = 0.38) and at an SFWBC of 50 × 10 3/μl or higher it increased considerably (iLR = ∞) (Table 2). CRP and the white blood cell count in the blood were statistically significantly increased in PJIs (P = 0.0002 and P = 0.01), but their AUCs showed lower values (0.768, 0.676). Discussion The erythrocyte sedimentation rate (ESR) and the C-reactive protein are highly sensitive serologic screening parameters for PJIs [25], but these parameters are also elevated in various chronic inflammatory conditions [26], after surgery, during postoperative complications and in other infectious diseases [27–29]. Spangehl et al reported a sensitivity of 96% and a specificity of 96% for CRP, but they excluded patients with inflammatory-tissue diseases, thus the specificity is likely lower in daily clinical work. In accordance to this, reported sensitivities and specificities of CRP levels in serum range in other studies from 71% to 94% and from 71% to 86%, respectively, using for CRP a threshold of 1 mg/dl [6,26,30]. The sensitivity and the specificity of the ESR ranged from 64% to 100% and from 56% to 87%, respectively, using an ESR cutoff-value of 30 mm/h [6,26,30]. Two further studies investigated the role of serum markers for predicting septic joints and revealed for CRP (cutoff-value 1 mg/dl) specificities of only 15% and 39% [31,32]. The examination of CRP levels showed
M. Lenski, M.A. Scherer / The Journal of Arthroplasty 29 (2014) 1105–1109
1107
Fig. 1. Receiver Operating Characteristic curves (ROCs). ROCs with the corresponding Area under the curve (AUC) of various inflammatory markers of patients with periprosthetic joint infection or arthralgia after total joint arthroplasty. IL-6 in the synovial fluid showed excellent diagnostic potential.
fair diagnostic potential in this study (AUC = 0.768). It is concluded that the sensitive screening parameters CRP and ESR have a relatively high number of false-positive test results and show fair diagnostic accuracy [6,33]. Serum CRP is used as screening parameter, thus we recommend using a threshold of 0.5 mg/dl to achieve a higher sensitivity and overlook fewer patients with PJIs. In order to find a serum parameter with higher accuracy, two studies documented significantly higher serum IL-6 levels in PJIs than in aseptic failure of prosthesis [26,30]. Di Cesare et al revealed that serum IL-6 levels are more accurate for predicting PJI than the ESR, CRP and pWBC. But serum IL-6 is also likely to be elevated after various inflammatory stimuli and conditions, such as sepsis, trauma, arthritis, meningitis and major surgeries [26]. Given that systemic levels of IL-6 are increased during PJIs, it appears likely that local levels of interleukin-6 in the synovial fluid are
also increased. Deirmengian et al evaluated the accuracy of several inflammatory markers for predicting PJIs compared to aseptic failure of prosthesis. Synovial IL-6 was the best biomarker with a sensitivity, specificity and accuracy of 100%, 100% and 1.00, and a cutoff-value of 13,350 pg/ml was used [6]. Furthermore Nilsdotter-Augustinsson et al used a threshold of 10,000 pg/ml, resulting in a sensitivity of 69% and a specificity of 93% [30]. In addition to this Jacovides et al introduced a cutoff-value of 4270 pg/ml and reported an excellent diagnostic potential of synovial IL-6 for diagnosing PJIs (AUC = 0.95); the sensitivity, specificity and accuracy were 87.1%, 100% and 94.6% [7]. Gollwitzer et al showed that synovial IL-6 has a good diagnostic potential (AUC = 0.807), especially if combined with increased cathelicidin LL-37 levels (AUC = 0.895), an antimicrobial peptide [34]. The synovial inflammatory markers glucose, lactate, LDH and the SFWBC have shown accuracy and clinical relevance for diagnosing
Table 1 Inflammatory Markers in Periprosthetic Joint Infections. Number IL-6 (pg/ml) Glucose (mg/dL) Lactate (mmol/L) SFWBC (×103/μL) LDH (U/l) CRPserum (mg/dl) pWBC (×103/μL)
n n n n n n n
= = = = = = =
22 24 21 25 25 30 30
Mean
AUC
Cutoff
308,099 30.4 12.8 69.6 4918 13.4 11.1
0.959 0.829 0.844 0.807 0.779 0.768 0.676
30,750 44.0 8.3 23.0 1423 0.5 11.5
SE 90.9% 79.2% 71.4% 60.0% 84.0% 93.3% 40.0%
(72.2–97.5) (59.5–90.8) (50.0–86.2) (40.7–76.6) (65.4–93.6) (78.7–98.2) (24.6–57.7)
SP 94.7% 78.6% 88.0% 94.3% 62.1% 21.4% 83.9%
(75.4–99.1) (60.5–89.8) (70.0–95.8) (81.4–98.4) (44.0–7.3) (10.2–39.5) (67.4–92.9)
−LR
+LR 17.27 (2.55–116.9) 3.69 (1.77–7.73) 5.95 (1.99–17.80) 10.50 (2.63–41.87) 2.22 (1.35–3.64) 1.19 (0.96–1.47) 2.48 (0.99–6.19)
0.10 0.27 0.33 0.42 0.26 0.31 0.72
(0.03–0.36) (0.12–0.59) (0.16–0.65) (0.26–0.69) (0.10–0.66) (0.07–1.42) (0.51–1.00)
Values in brackets are the corresponding 95% confidence intervals. AUC = Area under the curve, SE = sensitivity, SP = specificity, +LR = positive Likelihood-Ratio, −LR = negative Likelihood-Ratio, IL-6 = interleukin-6, SFWBC = synovial fluid white blood cell count, pWBC = peripheral white blood cells.
1108
M. Lenski, M.A. Scherer / The Journal of Arthroplasty 29 (2014) 1105–1109
Table 2 Interval Likelihood-Ratios. IL-6syn (in pg/ml)
Interval Likelihood-Ratio
≥49,000 10,000–49,000 b10,000 Glucosesyn (in mg/dL) ≤30 30–90 N90 Lactatesyn (in mmol/L) ≥10 10–4.3 b4.3 SFWBC (in ×103/μL) ≥50 10–50 b10
13.82 0.62 0.08
IL-6syn = interleukin-6, blood cell count.
5.83 0.48 0.29 14.29 0.63 0 ∞ 0.62 0.38 syn
= in the synovial fluid, SFWBC = synovial fluid white
septic arthritis [20,21,35–43] and are often examined in case of suspicion of PJI. In clinical studies about the role of synovial IL-6 in predicting PJIs, the concentrations of classic parameters in the synovial fluid, such as glucose, lactate and SFWBC, were often not examined [7,30,34]. Therefore the accuracy of synovial IL-6 could not be compared to the accuracy of parameters in standard clinical practice. The idea of this study was to introduce synovial IL-6 examination in clinical routine, when a PJI was suspected. The diagnostic strength of synovial IL-6 in differing septic from aseptic arthralgia of prosthesis was compared to the standard synovial fluid parameters in daily clinical routine. In this clinical trial, the determination of synovial IL-6 levels was the only test with excellent diagnostic potential (AUC = 0.959) and was much more accurate than synovial lactate (AUC = 0.844), glucose (AUC = 0.829) and SFWBC (AUC = 0.807), which also showed a good diagnostic potential for predicting PJIs. Synovial IL-6 was determined in 22 patients with PJI and 14 of them had an extremely elevated IL-6 level above 100,000 pg/ml. IL-6 promotes osteoclast activation [44,45], consequently leading to bone destruction [46]. Thus the extremely increased synovial IL-6 concentrations in PJIs may play a major role in the development of focal periprosthetic osteolysis and septic loosening of the prosthesis. The allocation of synovial fluid IL-6 levels into three continuous intervals (b10,000 pg/ml, 10,000 to 49,000 pg/ml and ≥ 49,000 pg/ml) with corresponding interval Likelihood-Ratios (0.08, 0.62 and 13.81) enables physicians to estimate the posttest probability of PJI without the use of cutoff-values. After having obtained synovial fluid for microbiological testing and laboratory, we recommend the immediate planning of operative and medicinal therapy, if the IL-6 level is above the threshold of 30,750 pg/ml so that a PJI becomes very likely. It should be stressed that common hospital IL-6 assays are optimized to measure cytokine levels in serum and not in the synovial fluid. They may provide results that deviate from our results. That is why it is necessary that future research should use IL-6 assays, whose reliability and performance had been documented in prior synovial biomarker studies. The diagnostic potential of IL-6 in this study was much higher than the diagnostic potential of the SFWBC, which is the most frequently used inflammatory marker in the synovial fluid for predicting a PJI. The diagnostic potential (in form of the AUC) of the SFWBC for predicting acute, hematogenic and chronic PJIs has never been calculated before and it was detected that the SFWBC was a good test (AUC = 0.807, sensitivity = 60.0%, specificity = 94.3%) in this clinical trial. Cipriano et al reported in a study about serum and synovial fluid markers for diagnosing chronic periprosthetic infections in patients with inflammatory arthritis an AUC of
0.938 for the SFWBC. But there are three main differences between Cipriano et al and our study: firstly, our study had a retrospective study design, secondly we included patients with acute, hematogenic and chronic infections of prosthesis and thirdly, the control group of this trial consisted of patients with arthralgia after TJA and not of patients with aseptic failure of prosthesis. Levels of inflammatory markers in the synovial fluid may differ between arthralgia after TJA and aseptic failure of prosthesis. Further, it is possible that participants in our trial with arthralgia after TJA may have suffered from acute aseptic inflammatory arthritis in the prosthetic joint. It has been reported that the SFWBC is also elevated in prosthetic joints during some types of aseptic inflammatory arthritis, such as rheumatoid arthritis and gout [33,47,48]. These differences may have led to the different AUCs of the SFWBC in both studies. Although the SFWBC is the most frequently investigated synovial marker in PJIs, the cutoff-values that are proposed by the authors vary from 2/μl to 50 × 10 3/μl, the given sensitivities from 33% to 94% and the specificities from 84% to 99% [6,29,30,41,48–50]. By using our iLRs, physicians will be able to interpret test results of the SFWBC more accurately. This study revealed that lactate levels in the synovial fluid have a good diagnostic potential (AUC = 0.844) for differing between PJI and arthralgia after TJA, especially lactate levels ≥10 mmol/l lead to a substantial increase of the likelihood of PJI. Prior studies have also reported about the high diagnostic accuracy of elevated lactate levels for predicting septic arthritis [18,37–40]. Nilsdotter-Augustinsson et al found marginally no statistically significant difference (P = 0.04) of lactate levels in PJI and aseptic failure of prosthesis, but the median was higher for PJIs. In our trial lactate levels were statistically significantly increased (P = 0.001) during PJIs. Lactate levels lower than 4.3 mmol/l ruled out the diagnosis. Low synovial glucose levels also showed a good diagnostic potential (AUC = 0.829) for predicting a PJI, but physicians have to evaluate the results carefully: synovial glucose levels may be increased in patients with diabetes mellitus, post-prandial or lack of the recommended 6 h fast [36]. Low glucose levels are often observed in PJIs and septic arthritis [36], but physicians must be aware of the conditions that may lead to false-negative results. We acknowledge that this retrospective clinical study is limited by several factors. Firstly, the workgroup convened by the Musculoskeletal Infection Society introduced a new definition for PJIs [19], but there is still no gold standard for diagnosing a PJI. Our criteria are accepted and were also used in prior trials [47,49]. Nevertheless, our criteria could have led to a misclassification of patients of having or not having a PJI. In addition to this we tried to rule out a PJI in the control group during their stay at hospital, but we cannot exclude that some patients in the control group could have suffered from a chronic low-grade infection. We tried to prevent bias, especially selection bias, by including every patient that presented with arthralgia after TJA at our emergency unit, if he met with the inclusion criteria. However, we believe that we have captured the spectrum of biomarker levels in patients with PJIs and that our results are representative. Conclusions We conclude that the IL-6 concentration in the synovial fluid is the best inflammatory marker for predicting a PJI. Synovial IL-6 levels below 10,000 pg/ml and synovial lactate levels below 4.3 mmol/l almost rule out the diagnosis PJI. IL-6 levels ≥ 49,000, lactate concentrations ≥ 10.0 mmol/l and an SFWBC ≥ 50 × 10 3/μl make the diagnosis PJI very likely. For IL-6 levels between 10,000 pg/ml and 49,000 pg/ml, physicians should be vigilant about the synovial glucose, lactate and SFWBC and estimate the likelihood of PJI. As inflammatory markers in native septic arthritis miss excellent diagnostic strength, future clinical studies should try to investigate the role of synovial IL-6 for predicting native septic joints.
M. Lenski, M.A. Scherer / The Journal of Arthroplasty 29 (2014) 1105–1109
References 1. Kurtz S, et al. Projections of primary and revision hip and knee arthroplasty in the United States from 2005 to 2030. J Bone Joint Surg Am 2007;89(4):780. 2. Pulido L, et al. Periprosthetic joint infection: the incidence, timing, and predisposing factors. Clin Orthop Relat Res 2008;466(7):1710. 3. Bozic KJ, et al. Hospital resource utilization for primary and revision total hip arthroplasty. J Bone Joint Surg Am 2005;87(3):570. 4. Segawa H, et al. Infection after total knee arthroplasty. A retrospective study of the treatment of eighty-one infections. J Bone Joint Surg Am 1999;81(10):1434. 5. McGarry SA, et al. Surgical-site infection due to Staphylococcus aureus among elderly patients: mortality, duration of hospitalization, and cost. Infect Control Hosp Epidemiol 2004;25(6):461. 6. Deirmengian C, et al. Synovial fluid biomarkers for periprosthetic infection. Clin Orthop Relat Res 2010;468(8):2017. 7. Jacovides CL, et al. Molecular markers for diagnosis of periprosthetic joint infection. J Arthroplasty, 2011;26(6 Suppl):99–103 e1. 8. Kalogianni DP, et al. Dry reagent dipstick test combined with 23S rRNA PCR for molecular diagnosis of bacterial infection in arthroplasty. Anal Biochem 2007;361 (2):169. 9. Kobayashi N, et al. Molecular identification of bacteria from aseptically loose implants. Clin Orthop Relat Res 2008;466(7):1716. 10. Parvizi J, et al. Periprosthetic infection: what are the diagnostic challenges? J Bone Joint Surg Am 2006;88(Suppl 4):138. 11. Zimmerli W, Trampuz A, Ochsner PE. Prosthetic-joint infections. N Engl J Med 2004;351(16):1645. 12. Ghanem E, et al. The use of receiver operating characteristics analysis in determining erythrocyte sedimentation rate and C-reactive protein levels in diagnosing periprosthetic infection prior to revision total hip arthroplasty. Int J Infect Dis 2009;13(6):e444. 13. Kobayashi N, et al. The comparison of pyrosequencing molecular gram stain, culture, and conventional gram stain for diagnosing orthopaedic infections. J Orthop Res 2006;24(8):1641. 14. Della Valle CJ, et al. The role of intraoperative gram stain in revision total joint arthroplasty. J Arthroplasty 1999;14(4):500. 15. Ghanem E, et al. Periprosthetic infection: where do we stand with regard to gram stain? Acta Orthop 2009;80(1):37. 16. Logters T, et al. Diagnostic accuracy of neutrophil-derived circulating free DNA (cf-DNA/NETs) for septic arthritis. J Orthop Res 2009;27(11):1401. 17. Mathews CJ, et al. Management of septic arthritis: a systematic review. Ann Rheum Dis 2007;66(4):440. 18. Carpenter CR, et al. Evidence-based diagnostics: adult septic arthritis. Acad Emerg Med 2011;18(8):781. 19. Parvizi J, et al. New definition for periprosthetic joint infection: from the Workgroup of the Musculoskeletal Infection Society. Clin Orthop Relat Res 2011;469(11):2992. 20. Martinot M, et al. Diagnostic value of serum and synovial procalcitonin in acute arthritis: a prospective study of 42 patients. Clin Exp Rheumatol 2005;23(3):303. 21. Li SF, et al. Diagnostic utility of laboratory tests in septic arthritis. Emerg Med J 2007;24(2):75. 22. Bedair H, et al. The Mark Coventry Award: diagnosis of early postoperative TKA infection using synovial fluid analysis. Clin Orthop Relat Res 2011;469(1):34. 23. Gallagher EJ. Evidence-based emergency medicine/editorial. The problem with sensitivity and specificity. Ann Emerg Med 2003;42(2):298. 24. McGee S. Simplifying likelihood ratios. J Gen Intern Med 2002;17(8):646. 25. Aggarwal VK, Rasouli MR, Parvizi J. Periprosthetic joint infection: current concept. Indian J Orthop 2013;47(1):10.
1109
26. Di Cesare PE, et al. Serum interleukin-6 as a marker of periprosthetic infection following total hip and knee arthroplasty. J Bone Joint Surg Am 2005;87(9):1921. 27. Ellitsgaard N, et al. Changes in C-reactive protein and erythrocyte sedimentation rate after hip fractures. Int Orthop 1991;15(4):311. 28. Hjortdahl P, et al. C-reactive protein: a new rapid assay for managing infectious disease in primary health care. Scand J Prim Health Care 1991;9(1):3. 29. Bauer TW, et al. Diagnosis of periprosthetic infection. J Bone Joint Surg Am 2006; 88(4):869. 30. Nilsdotter-Augustinsson A, et al. Inflammatory response in 85 patients with loosened hip prostheses: a prospective study comparing inflammatory markers in patients with aseptic and septic prosthetic loosening. Acta Orthop 2007;78(5):629. 31. Fottner A, et al. Can serum procalcitonin help to differentiate between septic and nonseptic arthritis? Arthroscopy 2008;24(2):229. 32. Ernst AA, et al. Usefulness of CRP and ESR in predicting septic joints. South Med J 2010;103(6):522. 33. Lachiewicz PF, Rogers GD, Thomason HC. Aspiration of the hip joint before revision total hip arthroplasty. Clinical and laboratory factors influencing attainment of a positive culture. J Bone Joint Surg Am 1996;78(5):749. 34. Gollwitzer H, et al. Antimicrobial peptides and proinflammatory cytokines in periprosthetic joint infection. J Bone Joint Surg Am 2013;95(7):644. 35. Shmerling RH, et al. Synovial fluid tests. What should be ordered? JAMA 1990; 264(8):1009. 36. Soderquist B, et al. Bacterial or crystal-associated arthritis? Discriminating ability of serum inflammatory markers. Scand J Infect Dis 1998;30(6):591. 37. Brook I, et al. Synovial fluid lactic acid. A diagnostic aid in septic arthritis. Arthritis Rheum 1978;21(7):774. 38. Mossman SS, Coleman JM, Gow PJ. Synovial fluid lactic acid in septic arthritis. N Z Med J 1981;93(678):115. 39. Riordan T, Doyle D, Tabaqchali S. Synovial fluid lactic acid measurement in the diagnosis and management of septic arthritis. J Clin Pathol 1982;35(4):390. 40. Gratacos J, et al. D-Lactic acid in synovial fluid. A rapid diagnostic test for bacterial synovitis. J Rheumatol 1995;22(8):1504. 41. Trampuz A, et al. Synovial fluid leukocyte count and differential for the diagnosis of prosthetic knee infection. Am J Med 2004;117(8):556. 42. Jeng GW, et al. Measurement of synovial tumor necrosis factor-alpha in diagnosing emergency patients with bacterial arthritis. Am J Emerg Med 1997;15(7):626. 43. Kortekangas P, et al. Synovial fluid leukocytosis in bacterial arthritis vs. reactive arthritis and rheumatoid arthritis in the adult knee. Scand J Rheumatol 1992;21(6):283. 44. Sims NA, Walsh NC. GP130 cytokines and bone remodelling in health and disease. BMB Rep 2010;43(8):513. 45. Cronstein BN. Interleukin-6—a key mediator of systemic and local symptoms in rheumatoid arthritis. Bull NYU Hosp Jt Dis 2007;65(Suppl 1):S11. 46. Park JY, Pillinger MH. Interleukin-6 in the pathogenesis of rheumatoid arthritis. Bull NYU Hosp Jt Dis 2007;65(Suppl 1):S4. 47. Cipriano CA, et al. Serum and synovial fluid analysis for diagnosing chronic periprosthetic infection in patients with inflammatory arthritis. J Bone Joint Surg Am 2012;94(7):594. 48. Kersey R, Benjamin J, Marson B. White blood cell counts and differential in synovial fluid of aseptically failed total knee arthroplasty. J Arthroplasty 2000; 15(3):301. 49. Spangehl MJ, et al. Prospective analysis of preoperative and intraoperative investigations for the diagnosis of infection at the sites of two hundred and two revision total hip arthroplasties. J Bone Joint Surg Am 1999;81(5):672. 50. Mason JB, et al. The value of white blood cell counts before revision total knee arthroplasty. J Arthroplasty 2003;18(8):1038.
Contents lists available at ScienceDirect
The Journal of Arthroplasty journal homepage: www.arthroplastyjournal.org
Synovial IL-6 AS Inflammatory Marker in Periprosthetic Joint Infections Markus Lenski, Cand. med. a, b, Michael A. Scherer, Dr. med. a a Department of Orthopedics and Trauma Surgery, Klinikum Dachau, Academic Teaching Hospital of the Ludwig-Maximilians-University of Munich, 7 Krankenhausstraße 15, 85221 Dachau, Germany b Faculty of Medicine, Technical University of Munich, Munich, Germany
a r t i c l e
i n f o
Article history: Received 17 June 2013 Accepted 14 January 2014 Keywords: periprosthetic joint infection inflammatory marker interleukin-6 synovial fluid synovial fluid white blood cell count lactate
a b s t r a c t We analyzed serum and synovial biomarkers of 69 patients. 31 of them suffered from a periprosthetic joint infection (PJI) and 38 from aseptic arthralgia after total joint arthroplasty. We used Receiver-OperatingCharacteristic-curves to calculate the Area-under-the-curve (AUC), cutoff-values, positive (+LR), negative (− LR) and interval-Likelihood-Ratios (iLR) for predicting a PJI. The most significant parameter was synovial interleukin-6 (IL-6) (cutoff-value ≥ 30,750 pg/ml, AUC = 0.959, SE = 90.0%, SP = 94.7%, + LR = 17.27), followed by synovial lactate (cutoff-value ≥ 8.3 mmol/l, AUC = 0.844, SE = 71.4%, SP = 88.0%, + LR = 5.95), and synovial glucose (cutoff-value ≤ 44 mg/dl, AUC = 0.829, SE = 79.2%, SP = 78.6%, +LR = 3.69). IL-6 ≥30,750 pg/ml and lactate ≥ 10 mmol/l make a PJI very likely, IL-6 b10,000 pg/ml or lactate b 4.3 mmol/l makes a PJI very unlikely. If none of these thresholds are met, physicians should use the iLR of IL-6, glucose and lactate to estimate the likelihood of PJI. © 2014 Elsevier Inc. All rights reserved.
The demographic change in Europe and North America as well as the medical progress in the field of arthroplasty leads to an increasing number of total joint arthroplasties, and with them the number of periprosthetic joint infections (PJIs) increases [1]. After total joint arthroplasty, an infection of the prosthesis occurs in 0.7% [2] and is a serious complication leading to revision arthroplasty, multiple surgical interventions, long-lasting disability of the patient and poor functional outcome [3,4]. Revision total joint arthroplasties are responsible for longer mean length of hospitalization, higher complication rates, for elevated mortality, especially in elderly patients and for the concomitant effect of increased hospital costs [3,5]. On account of this the diagnosis of PJI has to be set up rapidly and reliably, thus today’s research in the field of PJI concentrates on biomarkers that allow diagnosing PJIs with high sensitivity, specificity and accuracy. Deirmengian et al [6] and Jacovides et al [7] evaluated the role of various biomarkers in the synovial fluid in differing PJI from aseptic failure of prosthesis. The currently used biomarkers in daily clinical routine are either highly sensitive and less specific (e.g. polymerase chain reaction [8,9], erythrocyte sedimentation rate, C-reactive Protein [CRP] and peripheral white blood cells [pWBC] [10–12]) or highly specific and less sensitive (e.g. gram stain [13–15], synovial fluid white blood cell count [SFWBC] [16–18], synovial glucose[18]) and some show low diagnostic accuracy (e.g. synovial LDH, synovial total protein [18]). Deirmengian et al [6] and Jacovides et al [7] The Conflict of Interest statement associated with this article can be found at http://dx.doi.org/10.1016/j.arth.2014.01.014. Reprint requests: Markus Lenski, Gerhard Hanke Weg 22, 85221 Dachau, Germany. http://dx.doi.org/10.1016/j.arth.2014.01.014 0883-5403/© 2014 Elsevier Inc. All rights reserved.
revealed that especially the inflammatory marker IL-6 seems to have high diagnostic potential and accuracy for diagnosing PJIs. The aim of this study was (1) to investigate the role of synovial IL-6 in diagnosing PJI in daily clinical routine and (2) to compare its diagnostic potential to standard inflammatory parameters in the serum and in the synovial fluid in suspected PJI. The hypothesis was that IL-6 would be more useful for differing between both pathologic conditions than standard inflammatory markers and would be the first inflammatory parameter with excellent diagnostic potential for predicting PJI, when compared to patients with aseptic arthralgia after total joint arthroplasty (TJA). Patients and Methods 69 Patients that had undergone primary total joint arthroplasty and presented to the emergency unit of our hospital with symptoms consistent with a PJI, such as pain, swelling, redness, warming, impaired range of motion and fever, were included in this study. On admission, blood samples were taken to define the concentration of the C-reactive Protein (CRP) and the peripheral white blood cell count (pWBC) in a routine laboratory. The suspected joint was aspirated under strictly aseptic conditions and the levels of lactate, glucose, LDH and the synovial fluid white blood cell count (SFWBC) were examined in a clinical routine laboratory. In addition, the levels of interleukin-6 (IL-6) in the synovial fluid were determined by using the IMMUNOLITE 1000® Immunoassay System (SIEMENS Healthcare, Germany). The commercially available assay was optimized for synovial fluid, but not especially altered to account for synovial fluid matrix effects. Several dilution steps were necessary in order to account for the extremely elevated synovial levels compared to
1106
M. Lenski, M.A. Scherer / The Journal of Arthroplasty 29 (2014) 1105–1109
serum levels of IL-6. All assays were carried out according to the manufacturer’s instructions. The diagnosis criteria used for PJI were modeled after the Musculoskeletal Infection Society (MSIS) [19]. These criteria included a communicating sinus tract with the prosthesis, or a microbiologic test of the synovial fluid showing a positive microbiological culture, or at least four of the following five criteria: presence of pus in the affected joint, histological proof of infection, CRP level N 10 mg/dl [20], pWBC N 11 × 10 3/μl [21] and SFWBC N 20 × 10 3/μl [19,20,22] (n = 31). The control group (n = 38) consisted of patients who complained about aseptic arthralgia after TJA and had a CRP of less than 10 mg/dl, an SFWBC of less than 20 × 10 3/μl [19,22] and at least one negative microbiological culture. Arthralgia is the leading symptom in PJIs. In patients with arthralgia after TJA an infection was ruled out during their stay in hospital. The Receiver-Operating-Characteristic-curve (ROC) and the corresponding Area-under-the-curve (AUC) were calculated for each inflammatory parameter. Parameters with an AUC N 0.9 have an excellent and those with an AUC N 0.8 a good diagnostic potential. Optimal cutoff-values were determined by maximizing sensitivity and specificity (Youden’s J statistic). Cutoff-values lead to false positive and false negative test-result, leading to a loss of information of the obtained data, which is problematic for diagnosing PJIs. This is the reason why Likelihood-Ratios (LRs) and interval Likelihood ratios (iLRs) were calculated. Furthermore LRs and iLRs are more useful than sensitivity and specificity for clinical bedside application [23]. LRs are used to estimate the change from pretest to posttest probability of a disease. An LR above 1 increases the posttest-probability of a disease and an LR lower 1 decreases it. To be more specific, an LR of 5 increases the posttest disease probability + 30% and an LR of 10 increases it + 45%, whereas an LR of 0.5 decreases the posttest probability − 15%, an LR of 0.1 decreases it − 45% [24]. These estimated changes of the pretest to posttest probability can only be used, if the pretest probability of a disease is between 10% and 90%. The positive Likelihood-Ratio (+ LR) gives the change from pretest to posttest probability for a positive test result, the negative LikelihoodRatio (− LR) gives the change for a negative test result. Interval Likelihood-Ratios are given on a continuous scale of biomarker levels and use more of the contained information of inflammatory markers than LRs and therefore they are more useful for the estimation of the disease probability as well as for clinical decision making. The whole statistical analysis was performed by using SPSS for Windows (SPSS 17.0, Inc., Chicago, IL). A statistics institution affirmed our statistical approach. Participants of this clinical trial presented to our emergency unit of the hospital from October 2009 until October 2012. This study was conducted in accordance to the Declaration of Helsinki. No approval of the Ethic Review Committee of the Technical University of Munich, Germany, was sought for this retrospective epidemiologic study without personal data, as all data were only obtained if medically indicated. No blood test or arthrocentesis was done for the sole purpose of this study. Diagnostic procedures and therapeutic treatment were not directly influenced. The established standard operating workflow was used in case of suspicion of septic arthritis, providing a solid basis for routine data collection. Results 69 Patients took part in this clinical trial, 31 of them suffered from a perisprosthetic joint infection. The isolated pathogens in the PJI were Staphylococcus aureus (n = 11), coagulase negative staphylococci (n = 4), β-hemolytic streptococci (n = 5), gram negative bacteria (n = 8), Finegoldia magna (n = 4) and Enterococcus faecalis (n = 1). In two cases of PJIs, the growth of two pathogens was observed. 20 knee prostheses, 10 hip prostheses and 1 elbow prosthesis were infected. Arthralgia after TJA was the diagnosis in 38 patients. The
most frequently affected joint was the knee (n = 27), followed by the hip (n = 9) and the elbow (n = 2). The mean concentration of IL-6 in the synovial fluid during a PJI was 308,099 pg/ml and was statistically significantly higher (P = 0.01) than during arthralgia after TJA with a mean concentration of 10,333 pg/ml. The AUC was 0.959 (Fig. 1) with an optimal threshold at 30,750 pg/ml (sensitivity = 90.9%, specificity = 94.7%, + LR = 17.27, − LR = 0.10) (Table 1). The calculation of the interval Likelihood-Ratio showed that IL-6 levels below 10,000 pg/ml drastically decreased the posttest-probability of PJI (iLR = 0.08), IL-6 levels above 49,000 pg/ml led to a substantial change of the posttest probability of PJI (iLR = 13.82) (Table 2). The mean glucose level in the synovial fluid of patients with PJI was 30.4 mg/dl and statistically significantly (P = 0.001) decreased compared to patients with arthralgia after TJA with a mean of 71.2 mg/dl. The AUC of synovial glucose was 0.829, revealing an optimal cutoff-value at 44 mg/dl (sensitivity = 79.2%, specificity = 78.6%, + LR = 3.69, − LR = 0.27). The iLR revealed that glucose levels above 90 mg/dl decrease the likelihood of PJI, whereas glucose levels below 30 mg/dl increased it (Table 2). The arithmetical mean of lactate levels in the synovial fluid during PJI was statistically significantly (P = 0.001) increased compared to levels during arthralgia after TJA with a mean of 12.8 mmol/l compared to 5.7 mmol/l. The determined AUC of lactate was 0.844, leading to an optimal threshold of 8.3 mmol/l (sensitivity = 71.4%, specificity = 88.0%, + LR = 5.95, − LR = 0.33). The posttest probability of PJI decreased considerably at lactate levels below 4.3 mmol/l (iLR = 0) and increased drastically at lactate levels of 10 mmol/l or higher (iLR = 14.29). LDH levels in the synovial fluid had an AUC of 0.805 and showed a mean of 4918 U/l during PJI, which was statistically significantly (P = 0.007) increased compared to the mean of 1832 U/l during arthralgia after TJA. The best cutoff-value was 1423 U/l (sensitivity = 84.0%, specificity = 72.0%, + LR = 3.00, − LR = 0.22). The AUC of the SFWBC was 0.807, the mean SFWBC was 69.6 × 10 3/μl during PJI and 6.3 × 10 3/μl during arthralgia after TJA. The arithmetical means of both groups differed statistically significantly (P = 0.000008). The best threshold was an SFWBC of 23.0 × 10 3/μl (sensitivity = 60.0%, specificity = 94.3%, + LR = 10.50, − LR = 0.42). By using a threshold of 50 × 10 3/μl, the sensitivity was 52% and the specificity became 100%. At an SFWBC below 10 × 10 3/μl the likelihood of PJI decreased (iLR = 0.38) and at an SFWBC of 50 × 10 3/μl or higher it increased considerably (iLR = ∞) (Table 2). CRP and the white blood cell count in the blood were statistically significantly increased in PJIs (P = 0.0002 and P = 0.01), but their AUCs showed lower values (0.768, 0.676). Discussion The erythrocyte sedimentation rate (ESR) and the C-reactive protein are highly sensitive serologic screening parameters for PJIs [25], but these parameters are also elevated in various chronic inflammatory conditions [26], after surgery, during postoperative complications and in other infectious diseases [27–29]. Spangehl et al reported a sensitivity of 96% and a specificity of 96% for CRP, but they excluded patients with inflammatory-tissue diseases, thus the specificity is likely lower in daily clinical work. In accordance to this, reported sensitivities and specificities of CRP levels in serum range in other studies from 71% to 94% and from 71% to 86%, respectively, using for CRP a threshold of 1 mg/dl [6,26,30]. The sensitivity and the specificity of the ESR ranged from 64% to 100% and from 56% to 87%, respectively, using an ESR cutoff-value of 30 mm/h [6,26,30]. Two further studies investigated the role of serum markers for predicting septic joints and revealed for CRP (cutoff-value 1 mg/dl) specificities of only 15% and 39% [31,32]. The examination of CRP levels showed
M. Lenski, M.A. Scherer / The Journal of Arthroplasty 29 (2014) 1105–1109
1107
Fig. 1. Receiver Operating Characteristic curves (ROCs). ROCs with the corresponding Area under the curve (AUC) of various inflammatory markers of patients with periprosthetic joint infection or arthralgia after total joint arthroplasty. IL-6 in the synovial fluid showed excellent diagnostic potential.
fair diagnostic potential in this study (AUC = 0.768). It is concluded that the sensitive screening parameters CRP and ESR have a relatively high number of false-positive test results and show fair diagnostic accuracy [6,33]. Serum CRP is used as screening parameter, thus we recommend using a threshold of 0.5 mg/dl to achieve a higher sensitivity and overlook fewer patients with PJIs. In order to find a serum parameter with higher accuracy, two studies documented significantly higher serum IL-6 levels in PJIs than in aseptic failure of prosthesis [26,30]. Di Cesare et al revealed that serum IL-6 levels are more accurate for predicting PJI than the ESR, CRP and pWBC. But serum IL-6 is also likely to be elevated after various inflammatory stimuli and conditions, such as sepsis, trauma, arthritis, meningitis and major surgeries [26]. Given that systemic levels of IL-6 are increased during PJIs, it appears likely that local levels of interleukin-6 in the synovial fluid are
also increased. Deirmengian et al evaluated the accuracy of several inflammatory markers for predicting PJIs compared to aseptic failure of prosthesis. Synovial IL-6 was the best biomarker with a sensitivity, specificity and accuracy of 100%, 100% and 1.00, and a cutoff-value of 13,350 pg/ml was used [6]. Furthermore Nilsdotter-Augustinsson et al used a threshold of 10,000 pg/ml, resulting in a sensitivity of 69% and a specificity of 93% [30]. In addition to this Jacovides et al introduced a cutoff-value of 4270 pg/ml and reported an excellent diagnostic potential of synovial IL-6 for diagnosing PJIs (AUC = 0.95); the sensitivity, specificity and accuracy were 87.1%, 100% and 94.6% [7]. Gollwitzer et al showed that synovial IL-6 has a good diagnostic potential (AUC = 0.807), especially if combined with increased cathelicidin LL-37 levels (AUC = 0.895), an antimicrobial peptide [34]. The synovial inflammatory markers glucose, lactate, LDH and the SFWBC have shown accuracy and clinical relevance for diagnosing
Table 1 Inflammatory Markers in Periprosthetic Joint Infections. Number IL-6 (pg/ml) Glucose (mg/dL) Lactate (mmol/L) SFWBC (×103/μL) LDH (U/l) CRPserum (mg/dl) pWBC (×103/μL)
n n n n n n n
= = = = = = =
22 24 21 25 25 30 30
Mean
AUC
Cutoff
308,099 30.4 12.8 69.6 4918 13.4 11.1
0.959 0.829 0.844 0.807 0.779 0.768 0.676
30,750 44.0 8.3 23.0 1423 0.5 11.5
SE 90.9% 79.2% 71.4% 60.0% 84.0% 93.3% 40.0%
(72.2–97.5) (59.5–90.8) (50.0–86.2) (40.7–76.6) (65.4–93.6) (78.7–98.2) (24.6–57.7)
SP 94.7% 78.6% 88.0% 94.3% 62.1% 21.4% 83.9%
(75.4–99.1) (60.5–89.8) (70.0–95.8) (81.4–98.4) (44.0–7.3) (10.2–39.5) (67.4–92.9)
−LR
+LR 17.27 (2.55–116.9) 3.69 (1.77–7.73) 5.95 (1.99–17.80) 10.50 (2.63–41.87) 2.22 (1.35–3.64) 1.19 (0.96–1.47) 2.48 (0.99–6.19)
0.10 0.27 0.33 0.42 0.26 0.31 0.72
(0.03–0.36) (0.12–0.59) (0.16–0.65) (0.26–0.69) (0.10–0.66) (0.07–1.42) (0.51–1.00)
Values in brackets are the corresponding 95% confidence intervals. AUC = Area under the curve, SE = sensitivity, SP = specificity, +LR = positive Likelihood-Ratio, −LR = negative Likelihood-Ratio, IL-6 = interleukin-6, SFWBC = synovial fluid white blood cell count, pWBC = peripheral white blood cells.
1108
M. Lenski, M.A. Scherer / The Journal of Arthroplasty 29 (2014) 1105–1109
Table 2 Interval Likelihood-Ratios. IL-6syn (in pg/ml)
Interval Likelihood-Ratio
≥49,000 10,000–49,000 b10,000 Glucosesyn (in mg/dL) ≤30 30–90 N90 Lactatesyn (in mmol/L) ≥10 10–4.3 b4.3 SFWBC (in ×103/μL) ≥50 10–50 b10
13.82 0.62 0.08
IL-6syn = interleukin-6, blood cell count.
5.83 0.48 0.29 14.29 0.63 0 ∞ 0.62 0.38 syn
= in the synovial fluid, SFWBC = synovial fluid white
septic arthritis [20,21,35–43] and are often examined in case of suspicion of PJI. In clinical studies about the role of synovial IL-6 in predicting PJIs, the concentrations of classic parameters in the synovial fluid, such as glucose, lactate and SFWBC, were often not examined [7,30,34]. Therefore the accuracy of synovial IL-6 could not be compared to the accuracy of parameters in standard clinical practice. The idea of this study was to introduce synovial IL-6 examination in clinical routine, when a PJI was suspected. The diagnostic strength of synovial IL-6 in differing septic from aseptic arthralgia of prosthesis was compared to the standard synovial fluid parameters in daily clinical routine. In this clinical trial, the determination of synovial IL-6 levels was the only test with excellent diagnostic potential (AUC = 0.959) and was much more accurate than synovial lactate (AUC = 0.844), glucose (AUC = 0.829) and SFWBC (AUC = 0.807), which also showed a good diagnostic potential for predicting PJIs. Synovial IL-6 was determined in 22 patients with PJI and 14 of them had an extremely elevated IL-6 level above 100,000 pg/ml. IL-6 promotes osteoclast activation [44,45], consequently leading to bone destruction [46]. Thus the extremely increased synovial IL-6 concentrations in PJIs may play a major role in the development of focal periprosthetic osteolysis and septic loosening of the prosthesis. The allocation of synovial fluid IL-6 levels into three continuous intervals (b10,000 pg/ml, 10,000 to 49,000 pg/ml and ≥ 49,000 pg/ml) with corresponding interval Likelihood-Ratios (0.08, 0.62 and 13.81) enables physicians to estimate the posttest probability of PJI without the use of cutoff-values. After having obtained synovial fluid for microbiological testing and laboratory, we recommend the immediate planning of operative and medicinal therapy, if the IL-6 level is above the threshold of 30,750 pg/ml so that a PJI becomes very likely. It should be stressed that common hospital IL-6 assays are optimized to measure cytokine levels in serum and not in the synovial fluid. They may provide results that deviate from our results. That is why it is necessary that future research should use IL-6 assays, whose reliability and performance had been documented in prior synovial biomarker studies. The diagnostic potential of IL-6 in this study was much higher than the diagnostic potential of the SFWBC, which is the most frequently used inflammatory marker in the synovial fluid for predicting a PJI. The diagnostic potential (in form of the AUC) of the SFWBC for predicting acute, hematogenic and chronic PJIs has never been calculated before and it was detected that the SFWBC was a good test (AUC = 0.807, sensitivity = 60.0%, specificity = 94.3%) in this clinical trial. Cipriano et al reported in a study about serum and synovial fluid markers for diagnosing chronic periprosthetic infections in patients with inflammatory arthritis an AUC of
0.938 for the SFWBC. But there are three main differences between Cipriano et al and our study: firstly, our study had a retrospective study design, secondly we included patients with acute, hematogenic and chronic infections of prosthesis and thirdly, the control group of this trial consisted of patients with arthralgia after TJA and not of patients with aseptic failure of prosthesis. Levels of inflammatory markers in the synovial fluid may differ between arthralgia after TJA and aseptic failure of prosthesis. Further, it is possible that participants in our trial with arthralgia after TJA may have suffered from acute aseptic inflammatory arthritis in the prosthetic joint. It has been reported that the SFWBC is also elevated in prosthetic joints during some types of aseptic inflammatory arthritis, such as rheumatoid arthritis and gout [33,47,48]. These differences may have led to the different AUCs of the SFWBC in both studies. Although the SFWBC is the most frequently investigated synovial marker in PJIs, the cutoff-values that are proposed by the authors vary from 2/μl to 50 × 10 3/μl, the given sensitivities from 33% to 94% and the specificities from 84% to 99% [6,29,30,41,48–50]. By using our iLRs, physicians will be able to interpret test results of the SFWBC more accurately. This study revealed that lactate levels in the synovial fluid have a good diagnostic potential (AUC = 0.844) for differing between PJI and arthralgia after TJA, especially lactate levels ≥10 mmol/l lead to a substantial increase of the likelihood of PJI. Prior studies have also reported about the high diagnostic accuracy of elevated lactate levels for predicting septic arthritis [18,37–40]. Nilsdotter-Augustinsson et al found marginally no statistically significant difference (P = 0.04) of lactate levels in PJI and aseptic failure of prosthesis, but the median was higher for PJIs. In our trial lactate levels were statistically significantly increased (P = 0.001) during PJIs. Lactate levels lower than 4.3 mmol/l ruled out the diagnosis. Low synovial glucose levels also showed a good diagnostic potential (AUC = 0.829) for predicting a PJI, but physicians have to evaluate the results carefully: synovial glucose levels may be increased in patients with diabetes mellitus, post-prandial or lack of the recommended 6 h fast [36]. Low glucose levels are often observed in PJIs and septic arthritis [36], but physicians must be aware of the conditions that may lead to false-negative results. We acknowledge that this retrospective clinical study is limited by several factors. Firstly, the workgroup convened by the Musculoskeletal Infection Society introduced a new definition for PJIs [19], but there is still no gold standard for diagnosing a PJI. Our criteria are accepted and were also used in prior trials [47,49]. Nevertheless, our criteria could have led to a misclassification of patients of having or not having a PJI. In addition to this we tried to rule out a PJI in the control group during their stay at hospital, but we cannot exclude that some patients in the control group could have suffered from a chronic low-grade infection. We tried to prevent bias, especially selection bias, by including every patient that presented with arthralgia after TJA at our emergency unit, if he met with the inclusion criteria. However, we believe that we have captured the spectrum of biomarker levels in patients with PJIs and that our results are representative. Conclusions We conclude that the IL-6 concentration in the synovial fluid is the best inflammatory marker for predicting a PJI. Synovial IL-6 levels below 10,000 pg/ml and synovial lactate levels below 4.3 mmol/l almost rule out the diagnosis PJI. IL-6 levels ≥ 49,000, lactate concentrations ≥ 10.0 mmol/l and an SFWBC ≥ 50 × 10 3/μl make the diagnosis PJI very likely. For IL-6 levels between 10,000 pg/ml and 49,000 pg/ml, physicians should be vigilant about the synovial glucose, lactate and SFWBC and estimate the likelihood of PJI. As inflammatory markers in native septic arthritis miss excellent diagnostic strength, future clinical studies should try to investigate the role of synovial IL-6 for predicting native septic joints.
M. Lenski, M.A. Scherer / The Journal of Arthroplasty 29 (2014) 1105–1109
References 1. Kurtz S, et al. Projections of primary and revision hip and knee arthroplasty in the United States from 2005 to 2030. J Bone Joint Surg Am 2007;89(4):780. 2. Pulido L, et al. Periprosthetic joint infection: the incidence, timing, and predisposing factors. Clin Orthop Relat Res 2008;466(7):1710. 3. Bozic KJ, et al. Hospital resource utilization for primary and revision total hip arthroplasty. J Bone Joint Surg Am 2005;87(3):570. 4. Segawa H, et al. Infection after total knee arthroplasty. A retrospective study of the treatment of eighty-one infections. J Bone Joint Surg Am 1999;81(10):1434. 5. McGarry SA, et al. Surgical-site infection due to Staphylococcus aureus among elderly patients: mortality, duration of hospitalization, and cost. Infect Control Hosp Epidemiol 2004;25(6):461. 6. Deirmengian C, et al. Synovial fluid biomarkers for periprosthetic infection. Clin Orthop Relat Res 2010;468(8):2017. 7. Jacovides CL, et al. Molecular markers for diagnosis of periprosthetic joint infection. J Arthroplasty, 2011;26(6 Suppl):99–103 e1. 8. Kalogianni DP, et al. Dry reagent dipstick test combined with 23S rRNA PCR for molecular diagnosis of bacterial infection in arthroplasty. Anal Biochem 2007;361 (2):169. 9. Kobayashi N, et al. Molecular identification of bacteria from aseptically loose implants. Clin Orthop Relat Res 2008;466(7):1716. 10. Parvizi J, et al. Periprosthetic infection: what are the diagnostic challenges? J Bone Joint Surg Am 2006;88(Suppl 4):138. 11. Zimmerli W, Trampuz A, Ochsner PE. Prosthetic-joint infections. N Engl J Med 2004;351(16):1645. 12. Ghanem E, et al. The use of receiver operating characteristics analysis in determining erythrocyte sedimentation rate and C-reactive protein levels in diagnosing periprosthetic infection prior to revision total hip arthroplasty. Int J Infect Dis 2009;13(6):e444. 13. Kobayashi N, et al. The comparison of pyrosequencing molecular gram stain, culture, and conventional gram stain for diagnosing orthopaedic infections. J Orthop Res 2006;24(8):1641. 14. Della Valle CJ, et al. The role of intraoperative gram stain in revision total joint arthroplasty. J Arthroplasty 1999;14(4):500. 15. Ghanem E, et al. Periprosthetic infection: where do we stand with regard to gram stain? Acta Orthop 2009;80(1):37. 16. Logters T, et al. Diagnostic accuracy of neutrophil-derived circulating free DNA (cf-DNA/NETs) for septic arthritis. J Orthop Res 2009;27(11):1401. 17. Mathews CJ, et al. Management of septic arthritis: a systematic review. Ann Rheum Dis 2007;66(4):440. 18. Carpenter CR, et al. Evidence-based diagnostics: adult septic arthritis. Acad Emerg Med 2011;18(8):781. 19. Parvizi J, et al. New definition for periprosthetic joint infection: from the Workgroup of the Musculoskeletal Infection Society. Clin Orthop Relat Res 2011;469(11):2992. 20. Martinot M, et al. Diagnostic value of serum and synovial procalcitonin in acute arthritis: a prospective study of 42 patients. Clin Exp Rheumatol 2005;23(3):303. 21. Li SF, et al. Diagnostic utility of laboratory tests in septic arthritis. Emerg Med J 2007;24(2):75. 22. Bedair H, et al. The Mark Coventry Award: diagnosis of early postoperative TKA infection using synovial fluid analysis. Clin Orthop Relat Res 2011;469(1):34. 23. Gallagher EJ. Evidence-based emergency medicine/editorial. The problem with sensitivity and specificity. Ann Emerg Med 2003;42(2):298. 24. McGee S. Simplifying likelihood ratios. J Gen Intern Med 2002;17(8):646. 25. Aggarwal VK, Rasouli MR, Parvizi J. Periprosthetic joint infection: current concept. Indian J Orthop 2013;47(1):10.
1109
26. Di Cesare PE, et al. Serum interleukin-6 as a marker of periprosthetic infection following total hip and knee arthroplasty. J Bone Joint Surg Am 2005;87(9):1921. 27. Ellitsgaard N, et al. Changes in C-reactive protein and erythrocyte sedimentation rate after hip fractures. Int Orthop 1991;15(4):311. 28. Hjortdahl P, et al. C-reactive protein: a new rapid assay for managing infectious disease in primary health care. Scand J Prim Health Care 1991;9(1):3. 29. Bauer TW, et al. Diagnosis of periprosthetic infection. J Bone Joint Surg Am 2006; 88(4):869. 30. Nilsdotter-Augustinsson A, et al. Inflammatory response in 85 patients with loosened hip prostheses: a prospective study comparing inflammatory markers in patients with aseptic and septic prosthetic loosening. Acta Orthop 2007;78(5):629. 31. Fottner A, et al. Can serum procalcitonin help to differentiate between septic and nonseptic arthritis? Arthroscopy 2008;24(2):229. 32. Ernst AA, et al. Usefulness of CRP and ESR in predicting septic joints. South Med J 2010;103(6):522. 33. Lachiewicz PF, Rogers GD, Thomason HC. Aspiration of the hip joint before revision total hip arthroplasty. Clinical and laboratory factors influencing attainment of a positive culture. J Bone Joint Surg Am 1996;78(5):749. 34. Gollwitzer H, et al. Antimicrobial peptides and proinflammatory cytokines in periprosthetic joint infection. J Bone Joint Surg Am 2013;95(7):644. 35. Shmerling RH, et al. Synovial fluid tests. What should be ordered? JAMA 1990; 264(8):1009. 36. Soderquist B, et al. Bacterial or crystal-associated arthritis? Discriminating ability of serum inflammatory markers. Scand J Infect Dis 1998;30(6):591. 37. Brook I, et al. Synovial fluid lactic acid. A diagnostic aid in septic arthritis. Arthritis Rheum 1978;21(7):774. 38. Mossman SS, Coleman JM, Gow PJ. Synovial fluid lactic acid in septic arthritis. N Z Med J 1981;93(678):115. 39. Riordan T, Doyle D, Tabaqchali S. Synovial fluid lactic acid measurement in the diagnosis and management of septic arthritis. J Clin Pathol 1982;35(4):390. 40. Gratacos J, et al. D-Lactic acid in synovial fluid. A rapid diagnostic test for bacterial synovitis. J Rheumatol 1995;22(8):1504. 41. Trampuz A, et al. Synovial fluid leukocyte count and differential for the diagnosis of prosthetic knee infection. Am J Med 2004;117(8):556. 42. Jeng GW, et al. Measurement of synovial tumor necrosis factor-alpha in diagnosing emergency patients with bacterial arthritis. Am J Emerg Med 1997;15(7):626. 43. Kortekangas P, et al. Synovial fluid leukocytosis in bacterial arthritis vs. reactive arthritis and rheumatoid arthritis in the adult knee. Scand J Rheumatol 1992;21(6):283. 44. Sims NA, Walsh NC. GP130 cytokines and bone remodelling in health and disease. BMB Rep 2010;43(8):513. 45. Cronstein BN. Interleukin-6—a key mediator of systemic and local symptoms in rheumatoid arthritis. Bull NYU Hosp Jt Dis 2007;65(Suppl 1):S11. 46. Park JY, Pillinger MH. Interleukin-6 in the pathogenesis of rheumatoid arthritis. Bull NYU Hosp Jt Dis 2007;65(Suppl 1):S4. 47. Cipriano CA, et al. Serum and synovial fluid analysis for diagnosing chronic periprosthetic infection in patients with inflammatory arthritis. J Bone Joint Surg Am 2012;94(7):594. 48. Kersey R, Benjamin J, Marson B. White blood cell counts and differential in synovial fluid of aseptically failed total knee arthroplasty. J Arthroplasty 2000; 15(3):301. 49. Spangehl MJ, et al. Prospective analysis of preoperative and intraoperative investigations for the diagnosis of infection at the sites of two hundred and two revision total hip arthroplasties. J Bone Joint Surg Am 1999;81(5):672. 50. Mason JB, et al. The value of white blood cell counts before revision total knee arthroplasty. J Arthroplasty 2003;18(8):1038.
