The identification of serological markers to improve the diagnostic efficacy for the first stage of MSIS criteria was carried out [6, 8]. Studies show that in the postoperative period, the ESR and CRP levels are not always reliable to predict treatment failure [6, 22]. However, previous studies have paid the most attention to inflammatory and coagulation biomarkers, rarely focusing on pathogen-related indicators. In our study, we found that serum D-lactate and FIB have potential value for predicting PJI, especially serum D-lactate, with its high sensitivity, whereas the performances of D-dimer, FDP and other indicators (PT, INR, APTT) were poor.
In a previous study, SHARON et al [15, 23] first used D-lactate as a marker to distinguish bacterial infections in rat blood and human body fluids, including ascitic fluids, pleural fluids and cerebrospinal fluids. Then, Gratacos et al [24] demonstrated that D-lactate would be useful to differentiate bacterial synovitis from noninfectious arthritis, choosing a cutoff value of 0.05 mm/L, with a sensitivity and specificity of 85% and 96%. According to our data, serum D-lactate also had better sensitivity and NPV than ESR and FIB, suggesting D-lactate as a good screening test in situations in which the ESR and CRP may be falsely high or low. The 7 false-positive cases could not be explained definitively. Two reasons might account for this result. First, it might be influenced by hemoglobin because insufficient centrifugation and erythrolysis may cause false-positive D-lactate tests due to the similar absorbance wavelengths of 570 nm for D-lactate and 540 nm for hemoglobin [25]. Second, the complications of diabetes (2 cases), hypoproteinemia (1 case) and cervical cancer (1 case) might cause mild glucose metabolic disorders, which lead to pyruvic aldehyde increasing later and ultimately converting to D-lactate in trace amounts [26].
The FIB achieved similarly good performance in diagnosing PJI before revision arthroplasty. The sensitivity and specificity were 80.77% and 73.08%, respectively, when choosing 3.27 g/L, which was in line with previous studies in various research groups [27, 28]. Rui et al [12] reported that an FIB threshold of 4.01 g/L had a sensitivity of 76.3% and a specificity of 86.30%. Nevertheless, the sensitivity and NPV for predicting PJI of ESR and FIB were less prominent than those of serum D-lactate. Meanwhile, as FIB is routinely analyzed before surgery, the time when we would almost be able to confirm the diagnosis of PJI is not practical in first-line screening for PJI in outpatient wards.
FDP, produced when plasmin degrades fibrin and fibrinogen, is related to the activation of fibrinolytic systems. Although Wang et al [29] indicated that FDP analysis would provide diagnostic support for PJI, our result (AUC: 0.69) indicated that the value of FDP for diagnosing PJI is poor. However, the wax and wane with fluctuations of sensitivity and specificity in the predictive cutoff values may be influenced by the sample sizes, genus and species of organisms, patient selection, study biases and clinical laboratory standardization [30].
Regarding D-dimer, we obtained the opposite outcomes (sensitivity 76.00%, specificity 60.00%, PPV% 65.52 and NPV 69.57%) of plasma D-dimer compared with the finding of Shahi, who insisted that serum D-dimer had great value (sensitivity 89.47%, specificity 92.75% PPV% 83.61% and 95.52%) in the assessment of the diagnosis of PJI, with only limited studies further validating this conclusion [31]. Although D-dimer is already one of the novel definitions of periprosthetic joint infection, a number of studies have demonstrated that D-dimer is of limited value for diagnosing PJI [8, 27, 28]. Two factors might contribute to the appearance of substantial discrepancies. First, D-dimer was detected using plasma instead of serum in most medical facilities, while Shahi and colleagues chose serum to determine the level of D-dimer [32]. The difference between plasma and serum may play a role. Second, our patients were mostly Asian, rather than patients in the American population, which is predominantly Caucasian and African American. Different levels of D-dimer in diverse races have been reported, such as between African American and Caucasian patients, which may contribute to the discrepancy from the studies by Shahi et al. [31, 33] and Parvizi et al.[8] Due to the discrepancy, D-dimer as an indicator of PJI should still be regarded with caution, and further studies should be conducted to identify the value of D-dimer for predicting PJI.
Recently, Katsiaryna et al [25] found that the levels of synovial D-lactate were significantly higher in patients in the PJI group than in those in the aseptic group. Choosing 1.26 mmol/l as the optimal threshold value, the sensitivity of 86.40% of synovial D-lactate outperformed both the SF WBC count and SF PMN% (79.5% and 56.8%, respectively), which was mostly consistent with our studies indicating that selecting 1.56 mmol/L as an optimal predictive cutoff resulted in a sensitivity of 95.65%, compared with the sensitivities of 70.00% and 80.00% for SF WBC count and SF PMN%, respectively. Although synovial D-lactate has better sensitivity than serum D-lactate (95.65% and 88.46%, respectively), it was weaker in specificity, PPV and NPV (68.00%, 73.33% and 81.82%) and (73.08%, 76.70% and 86.40%), respectively. However, when synovial D-lactate fails to diagnose or exclude PJI, there are still other biomarkers in synovial fluid to evaluate, such as SF WBC count, SF PMN%, leucocyte esterase, and interleukin 6, in contrast to the few serological options available. Obtaining synovial fluid is an invasive procedure that not only causes pain but also increases the risk of introducing infection into the joint. In addition, it is not always easy and smooth to aspirate enough synovial fluid, especially in the hip [34]. In summary, synovial D-lactate, the assessment of which involves a sophisticated, invasive process and an increased possibility of joint infection, has limited value as an extensive outpatient screening test.
There are several limitations to this study. First, only a limited number of patients were included from our medical center in the cohort analysis. Hence, the conclusions drawn should be further verified in larger samples and multiple centers. Second, the serum samples were all obtained before revision after antimicrobial therapy; thus, the influence of antibiotics on D-lactate cannot be ruled out. Third, the major limitation is that the increased levels of blood glucose, pyruvic aldehyde and ketone bodies in diabetic patients as well as the hemoglobin, L-lactate, fructose 1,6-diphosphate and glycerol 3-phosphate released after erythrocyte lysis might result in partially false-positive patients [26]. Hence, more efficient and accurate detection means and instruments should be chosen for further analysis.
In conclusion, serum D-lactate is a promising serological biomarker for predicting PJI, at least compared to ESR, FIB and synovial D-lactate. D-dimer and FDP may be of limited value in the diagnosis of PJI. The advantages of high sensitivity, pathogenic specificity, rapid availability of the results, little sample demand, concise and minimally invasive processes and low cost make serum D-lactate extremely useful as a real-time screening test for PJI. The combination with confirmatory synovial fluid biomarkers with higher specificity or improvements in the detection method might potentially improve its specificity.