Diagnostic usefulness of sCD 163 , procalcitonin and neopterin for sepsis risk assessment in critically ill patients

Background. Sepsis is one of the most common causes of hospitalization and it is characterized by a high mortality rate in spite of the great progress in diagnosis and treatment achieved in recent years. Early diagnosis of sepsis is one of the most important elements of effective treatment. The clinical symptoms are not specific and biomarkers are considered to be useful tools in sepsis diagnostics.

Sepsis is a serious medical condition characterized by a high mortality rate in spite of the great progress in diagnosis and treatment achieved in recent years. 1 Annual mortality due to sepsis ranges from 30 to 50 cases per 100,000 people and thus sepsis ranks in the top 10 causes of death. 1,2The mortality rate corresponds to the severity of the disease and varies from 25-30% in severe sepsis to as high as 40-70% in septic shock. 2 Early diagnosis of sepsis is one of the most important elements of effective treatment. 3Traditional clinical signs and symptoms of inflammation and laboratory parameters, i.e. body temperature, leukocyte count, erythrocyte sedimentation rate (ESR) or even C-reactive protein elevation, have been found to have low diagnostic accuracy. 1,4any biomarkers have been introduced in the diagnosis and treatment monitoring of sepsis, but their clinical utility has not yet been clearly determined. 1,3Procalcitonin (PCT) has been used commonly, as its diagnostic value is relatively well described, although it is not an ideal sepsis marker. 5Another marker of potential value in the diagnosis of infection is neopterin (NPT), however its value in sepsis has not yet been sufficiently evaluated and NPT determination has not been introduced as a routine parameter. 4According to the current publications about soluble CD163 (sCD163), one can conclude that it is a marker which correlates with common markers of sepsis.It can be used as an adjunct to the diagnostic process.][8] PCT is synthesized by thyroid gland C cells under physiological circumstances, whereas inflammation stimulates other cells to produce this polypeptide.The PCT level starts to rise within 6 h after an inflammation process is initiated, reaches maximum values between 12 and 48 h, and then returns to normal values if the infection agent has been eradicated. 5A persisting high or continuously increasing concentration of PCT suggests a systemic inflammatory reaction or septic complication after surgery.PCT concentration remains low in local infections without systemic response.There are many reports indicating that PCT is preferably induced in patients with sepsis of bacterial origin, especially severe sepsis and septic shock.Moreover, PCT concentration often correlates with the severity of the inflammation.One of the limitations of this parameter is the nonspecific increase in many clinical states regardless of sepsis, including surgical procedures, injuries, heat shock, burns, long-lasting cardiogenic shock and severe systemic inflammation process. 9Considering these facts as well as the lack of optimal cut-off values for PCT for different groups of patients, results should be interpreted cautiously in a sepsis risk assessment. 10eopterin is excreted by monocytes/macrophages mostly in response to interferon gamma stimulation, whereas IL-4 (interleukin -IL), IL-10 and IL-12 inhibit NPT excretion. 11,12Neopterin is an early and sensitive marker of cellular response but it lacks specificity.It is considered to be a useful tool in the investigation of clinical cases in which cellular response is crucial.Elevation of NPT often precedes the clinical manifestation of inflammation or seroconversion. 13The correlation between NPT serum level and the risk of septic complications has been described. 5Localized infections usually do not increase NPT concentration, whereas the process of persistent inflammation and systemic infection with the activation of cellular response is likely to increase NPT level. 12Sepsis is the most common cause of NPT elevation; moreover, higher NPT concentrations correlate with poor prognosis for patients. 14D163 (cluster of differentiation -CD) is a glycoprotein present only on the cell membrane on the surface of macrophages and monocytes.It is a member of the scavenger receptor cysteine-rich (SRCR) family class B. 15,16 CD163 is a 130 kDa protein consisting of 9 extracellular SRCR protein domains which are linked to a short transmembrane segment and a short cytoplasmic tail.In vitro tests have revealed that glucocorticoids, IL-10 and IL-6 increase the amount of CD163, while IL-4 and IL-13 do not raise the level of CD163 concentration, and TNF-α (tumor necrosis factor alpha), IFN-γ (interferon gamma) and CXCL4 (CXC chemokine ligand 4) decrease the concentration of CD163.Such variable responses to pro-and anti-inflammatory mediators suggest that CD163 is expressed mostly on M2 macrophages.M2 macrophages participate in tissue and wound repair, angiogenesis and tumor progression.CD163 is not only present as a membrane-bound protein, but also circulates in blood as a soluble CD163 (sCD163) and it is actively released from the cell membrane of monocytes and macrophages. 16sCD163 is present in serum and other body fluids. 15In vitro studies have also shown that sCD163 secretion by monocytes and macrophages might be induced by TLR (toll-like receptor) or LPS (lipopolysaccharide). Shedding of sCD163 is equal to TNF-α response to LPS.Oxidative stress also leads to significant shedding of CD163.Even without inflammatory stimulation, small amounts of sCD163 are released to circulation. 16CD163 has been identified as a hemoglobin scavenger receptor (HbSR).The receptor is capable of capturing haptoglobin-hemoglobin complexes (Hp-Hb) formed during intravascular hemolysis.HbSR and its ligands are directly involved in the inflammatory response. 17emoglobin and free heme play an important role in the occurrence of sepsis.Pathological agents like hemolysis, DIC and ischemic reperfusion, as well as the toxicity and side effects of drugs may lead to a release of hemoglobin and heme.At the same time, oxidation and reduction reactions catalyzed by hydrogen peroxide lead to a release of oxygen free radicals which results in organ damage.It is necessary to remove the hemoglobin and free heme in a short time.Haptoglobin (Hp) takes part in the removal of hemoglobin, and thus it inhibits the reactions of oxidative damage and is involved in anti-inflammatory response.Hemoglobin can be removed by phagocytic cells only when it is joined in a complex with haptoglobin (Hp-Hb) and through the CD163. 18CD163 is a specific biomarker of macrophages used in many clinical conditions such as coronary artery disease, atherosclerosis, rheumatoid arthritis, cancer and multiple sclerosis.It also plays an important role in the detection of inflammation, sepsis, bacteremia, mononucleosis, Crohn's disease, celiac disease, hemophagocytic syndrome, macrophage activation syndrome, acute liver failure, burns, preeclampsia in pregnancy and many other conditions. 15,16In numerous of these diseases, patients are exposed to large amounts of free heme due to intravascular hemolysis and tissue damage.CD163 is highly efficient in the removal of potentially toxic free hemoglobin from the circulation and tissues with inflammation, so sCD163 is to be considered a prognostic marker in many diseases, especially inflammatory ones.
Taking into consideration the advantages and limitations of PCT and NPT, we decided to determine the clinical value of serum measurements of sCD163, PCT and NPT as indicators of sepsis in critically-ill patients, that are at extreme risk of sepsis.Furthermore, the aim of the study was to compare their usefulness using ROC (receiver operator characteristic) curve analysis in this specific group of patients and to calculate cut-off values, exceeding which indicates elevated risk of sepsis.

Patients
Fifty-four vein blood samples were drawn from 30 patients of the Anesthesiology and Intensive Care Unit (AICU) of the Wroclaw Medical University Hospital.Samples were collected twice from each of 24 patients, on the 1 st and 3 rd day of hospitalization in the AICU.Only one sample was collected from 6 patients, because 4 patients died within 3 days and the remaining 2 were transferred to another hospital.Patient characteristics are presented in Table 1.All patients were divided into 2 groups -a study and a reference group.The study group consisted of 16 patients with an established diagnosis of sepsis at the time of sample collection.The reference group consisted of patients in a severe general condition, who didn't meet the diagnostic criteria of sepsis at the moment of material collection.This group included 14 patients.

Material
As material for this study we used serum remaining from samples collected for routine laboratory tests from patients of the AICU.All serum samples were frozen at -20°C and were stored at this temperature until the assay except for PCT which was performed immediately as a routine test.Use of this material was approved by the Bioethics Committee of Wroclaw Medical University.

Measurement of sCD163 concentration
The CD163 Quantikine ELISA Kit (R & D Systems, Minneapolis, USA) was used to measure the concentration of sCD163 in serum.The procedure was performed according to the manufacturer's instructions and the samples were diluted 10-fold according to the manufacturer's recommendations.

Measurement of neopterin concentration
A Neopterin ELISA kit (IBL International GMBH, Hamburg, Germany) was used to measure the concentration of neopterin in serum.The procedure was performed according to the manufacturer's instructions.

Procalcitonin concentration
The concentration of PCT in the serum samples was determined as a routine test in all patients by the University Center of Laboratory Diagnostics in Wroclaw.The measurements were performed via the automatic VIDAS ® B.R.A.H.M.S. PCT assay (BioMérieux, Marcy l'Etoile, France).

Statistical analysis
MedCalc v.14.12.0 (Ostend, Belgium) was used for the statistical calculations.The data was described with mean, highest and lowest value, standard deviation, me- dian and D' Agostino-Pearson test for normal distribution.
A Mann-Whitney rank-sum test for independent samples was used for the comparison between groups divided by the presence or absence of sepsis.A Wilcoxon test for paired samples was used to compare the values obtained on the 1 st and 3 rd day.A p value < 0.05 was considered statistically significant.The ROC curves were prepared for PCT, sCD163 and NPT and the areas under curves were compared using the method of DeLong et al. available in the MedCalc software.

Results
Table 2 shows the summarized results for the serum procalcitonin, sCD163 and neopterin measurements in patients, including the division into two groups.

Comparison of results between groups
We divided the results of the PCT, sCD163 and NPT measurements into two groups depending on the day of material collection (1 st and 3 rd day of hospitalization) and according to the clinical condition of the patients (with or without diagnosis of sepsis at the moment of sample collection -study and reference groups).Using the Wilcoxon test, we found that the differences between the 1 st and 3 rd day of material collection for PCT (p = 0.1005), sCD163 (p = 0.5129) and NPT (p = 0.4576) were not statistically significant.In contrast, there were statistically significant differences in PCT (p < 0.0001), sCD163 (p = 0.0001) and NPT (p = 0.0001) concentration between the study and reference groups, which was confirmed by a Mann-Whitney test.In Figs. 1, 2, 3, we show a graphical comparison between the study and reference groups including all the results obtained, regardless of the day of sample collection.

ROC curves
Considering that there were no differences found in PCT, sCD163 and NPT concentrations between the groups divided on the basis of the day of hospitalization and the total group of patients was relatively small, the  Comparison of the areas under the curves (AUC) for PCT, sCD163 and NPT revealed that there was no statistically significant difference, so it could be assumed that the PCT, sCD163 and NPT have similar diagnostic value in assessing the risk of sepsis development in ICU patients.The parameters describing the ROC curves and AUC comparison data are shown in Tables 3 and 4.
The most sensitive parameter was the NPT (100%), but at the same time it had the lowest diagnostic specificity (58.3%).PCT is more specific but less sensitive than NTP, whereas sCD163 is characterized by the highest diagnostic sensitivity and specificity in our study group.

Discussion
Early diagnosis of sepsis enables appropriate treatment and it is essential in emergency and intensive care units.Biomarkers are considered to be useful tools in sepsis diagnostics as they make it possible to differentiate the infection etiology and assess the severity of the inflammation process. 3Many studies have proved that procalcitonin is the most valuable biomarker, characterized by relatively high diagnostic accuracy, but one must remember the numerous limitations of this parameter as a sepsis marker. 10nother parameter is neopterin, which is a marker of cellular activation in inflammation and its concentration is related to the risk of sepsis. 5In spite of many clinical studies performed on NPT's diagnostic value, including as a predictor of mortality in sepsis, it still has not been determined what the total benefit is of NPT determination for patients and health care units.For these reasons, NPT is not a routinely used parameter.Over the last several  years, soluble CD163 has been studied by many groups of scientists as a potential marker for the diagnosis of sepsis.sCD163 concentration is not routinely measured in patients with suspicion of sepsis and there is still a need for a determination of its clinical utility.The aim of our study was the evaluation of the clinical value of sCD163 as a marker of sepsis and a comparison of it with procalcitonin and neopterin in a very specific group of ICU patients using a comparison of ROC curves and the determination of optimal cut-off values.To reliably evaluate the diagnostic value of sCD163 in the diagnosis of sepsis, the study group consisted of patients with similar clinical conditions and demographics factors to the reference group.Patients in both groups were in severe general condition, although the reference group did not meet the criteria of sepsis at the moment of sample collection.
An elevated concentration of all markers was observed in both groups in comparison to the cut-off values suggested by the manufacturers of the tests.Yet there was a significant difference between the study and reference groups for PCT, sCD163 and NPT, which is shown in Figs. 1, 2, 3 (sCD163 median in the reference group was 0.586 mg/L vs. 1.257 mg/L in the study group, p = 0.0001; for PCT 0.61 vs. 14.95 µg/L, respectively, p < 0.0001; for NPT 16.15 nmoL/L vs. 68.14nmoL/L, respectively, p = 0.0001).
There were similar observations made of PCT by Meisner et al.Their results confirmed the elevation of PCT in multi-organ dysfunction syndrome (MODS) and sepsis.They observed a correlation with the SOFA scale, and usually a higher concentrations of PCT in patients that died due to MODS or sepsis than in survivors. 19Also, Sudhir et al. have confirmed the increase of the SOFA result with increasing PCT concentration, but their study revealed no correlation between PCT and the severity of sepsis or mortality rate. 20eopterin's usefulness in sepsis was evaluated by Ploder et al. in a study that included patients after severe trauma and patients with sepsis, and they confirmed the elevation of NPT levels in these patients. 21In a retrospective study by Fisgin et al., NPT was assessed as a marker of mortality in patients with sepsis.Their results confirmed significantly higher concentrations of NPT in patients that died of sepsis (median 15 ng/mL) than in survivors (median 5 ng/mL, p = 0.03). 22Also, other clinical studies performed in ICU patients with sepsis, septic shock, MODS or infection have revealed that NPT makes it possible to distinguish between patients with confirmed infection and those without septic complications. 11,23ur results also show that the concentration of sCD163 was higher in the study than in the reference group.Median concentration in the study group was 1.257 mg/L and it was 2 times higher compared to the reference group (0.586 mg/L).The AUC value for sCD163 was 0.847.Based on a ROC curve analysis, the cut-off value was 1.146 mg/L with a diagnostic sensitivity of 82.1% and specificity 91.7%.Despite the relatively small group of patients, our results suggest a potential high diagnostic value of sCD163 measurements in sepsis diagnosis.
Feng et al. achieved similar results while measuring the concentration of sCD163 in the serum of 102 patients with sepsis and 30 patients with SIRS without infection in the aim of sepsis diagnosis, evaluation its severity and further prognosis.On the first day of hospitalization, the concentration of sCD163 in the septic patients was 2.51 times higher than in the patients with SIRS (p = 0.001).The authors noticed the maximum concentration on the third day of hospitalization, then a slow reduction in the following days.According to the authors, if the concentration doesn't decrease it means a worse prognosis for the patient.In our study there was no significant difference in concentration between the first and the third day of hospitalization, which could be the consequence of stabilizing the patients' condition.The ROC curve analysis by Feng et al. for the ability to discriminate between SIRS and sepsis showed that the area under the curve of sCD163 was 0.856.The cut-off value was 1.49 mg/L with a sensitivity of 74% and a specificity of 93.33%.In the present study we obtained higher diagnostic sensitivity but slightly less diagnostic specificity for a similar cutoff value.The authors also evaluated the usefulness of sCD163 concentration for the diagnosis and prognosis of sepsis.Their results suggest that sCD163 has clinical value for the recognition and prognosis of sepsis.The researchers also showed that on the first day of hospitalization the concentration of sCD163 in the serum of septic patients was significantly higher than in the patients with SIRS; additionally it better indicates sepsis diagnosis than the routinely-used PCT and CRP.sCD163 could be a valuable parameter as a specific and sensitive marker of infection.sCD163 is also a good indicator of the severity of sepsis and its monitoring. 6lso, a similar level of increase of the concentration of sCD163 as in this study was observed by Møller et al., who assessed the value of this parameter in patients with pneumococcal bacteremia (positive blood cultures for the presence of Streptococcus pneumoniae).They received 1.7 times higher concentrations of sCD163 in the study group compared to healthy subjects (p < 0.0001).The concentration of sCD163 was significantly increased in septic patients (median 4.6 mg/L) compared to healthy persons (median 2.7 mg/L).The study also showed that there is no difference in the concentrations of sCD163 depending on gender.However there is an age-related difference because the concentrations of sCD163 in older patients was lower.This observation may be one of the reasons for the relatively lower concentrations of sCD163 obtained in this study, but due to the small size of the group, the analysis in the age subgroups was not credible.The ROC curve for predicting the risk of death made by Møller et al. was characterized by the AUC = 0.82 with sensitivity 82% and specificity 79% and it indicated a high diagnostic value for this parameter in the assessment of the risk of death. 24lightly less diagnostic value than in this study was found by Su et al. who, in their work, defined the diagnostic value of sCD163 based on research carried out in the ICU in 100 septic patients and 30 patients with SIRS.These authors obtained an AUC for sCD163 of 0.696.The concentration of sCD163 in the septic patients was significantly higher (2.52 times) than in the patients with SIRS (p < 0.001).Based on the ROC curve analysis, the specified cut-off value was 2.84 mg/L with a sensitivity of 53.5% and specificity of 78.9%.This cut-off value is higher than in our study, but it should be taken into consideration that the values obtained in ELISA are affected not only by the composition of the study group but also by assay characteristics i.e. antibodies specifity, linearity, limit of quantification, precision and accuracy.Moreover, Su et al. observed that the median of the concentrations of sCD163 increased during hospitalization in patients who did not survive, while in survivors there was a decreasing tendency.According to the authors, the concentration of sCD163 has a positive diagnostic value for the prognosis and assessment of the disease dynamics. 7efinitely a higher level of increase of the concentration of sCD163 was found by Kjaergaard et al., who observed changes in the concentration of sCD163 during a 4-day stay in a hospital ward in 21 septic patients compared to 15 patients without sepsis and a control group (healthy people).During the 4 days, the concentrations of sCD163 were higher (3.44 times) in patients with sepsis than in those without sepsis (p < 0.001).The AUC of the ROC curve prepared by the authors to distinguish sepsis was 0.95.The cut-off value was 1.74 mg/L with a high sensitivity and specificity, both equaled 93%.The results reported by Kjaergaard et al. showed the high diagnostic value of sCD163.In this study we also demonstrated statistically significant differences between the concentrations of sCD163 in the study group (patients with sepsis) and the reference group (patients without sepsis, in severe condi-tion), but the AUC (0.847), sensitivity and specificity was lower than in the Kjaergaard et al. publication. 25bdelrahman et al. likewise analyzed the concentrations of sCD163 in ICU patients, which were divided into a study group (patients with sepsis) and a reference group (patients with SIRS).Based on the results obtained, the authors checked the diagnostic value of sCD163 in the diagnosis of sepsis and in the assessment of its severity.The researchers received a similar level of increase of sCD163 concentration as in this study.The septic patients had an average of 2.9 times higher concentrations of sCD163 than the patients with SIRS (p < 0.001).The cut-off value obtained from the ROC curve analysis was 5.36 mg/L with a sensitivity of 93.5% and specificity of 90.2%, which confirmed the high diagnostic value of this parameter in sepsis diagnosis.The authors didn't specify the AUC value in their publication. 8 lower diagnostic value was found by Ingels et al., who determined the concentrations of sCD163 in seriously ill patients.They obtained 1.65 times higher values in comparison with healthy people (p < 0.0001).The authors observed a relationship between the presence of sepsis and the concentration of sCD163; patients with sepsis recognized at the moment of admission to the hospital had 1.29 times higher concentrations than patients without sepsis (p < 0.0001).The researchers didn't show in their work on the ROC curve for the differentiation of sepsis, only for predicting mortality (AUC = 0.68). 26efinitely different results were found by Gaïni et al. who, in 2006, assessed the concentrations of sCD163 in patients with a community-acquired infection.The reference group consisted of healthy people, and the median concentration of sCD163 in this group was 1.9 mg/L (range: 0.4-4.7 mg/L).The ROC curve analysis showed that the AUC was low, equal to 0.58, and there weren't statistically significant differences between the studied groups of patients.The authors concluded that the concentrations of sCD163 did not differentiate patients with the infection from those without the infection. 26In 2008, Gaïni et al. determined the concentrations of sCD163 in patients with sepsis and SIRS.In septic patients, the concentrations were higher than in the control group (p < 0.0001), however there wasn't a statistically-significant difference between patients with sepsis and SIRS. 28n the publications mentioned above, the authors determined the concentrations of sCD163 in serum samples, but this parameter is also very efficient when measured in urine.The study by Su et al. revealed that in septic patients the concentration was 2.34 times higher than in patients with SIRS (p < 0.001).In the reference group, the concentration was undetectable.The AUC of ROC prepared to distinguish SIRS from sepsis was 0.83 with a sensitivity of 82.5% and specificity of 75%, and the cutoff value was equal to 0.043 mg/L.Moreover, the results correlated with the concentrations of sCD163 in serum (r = 0.511; p < 0.001). 18In conclusion, the assumptions of this study were confirmed by the results obtained.Based on the above-mentioned publications and the results received, sCD163 is a promising parameter, however its clinical utility has not yet been adequately determined.Similarly, neopterin, despite a relatively high sensitivity, hasn't been introduced as a routine laboratory parameter.In some cases, measurement of the concentration of NPT and in particular sCD163 provides valuable information for diagnosis, as shown in the above-mentioned reports, so there is a necessity to investigate these parameters and specifically determine their diagnostic usefulness.Both markers could be an additional parameters to PCT in order to make diagnosis easier and accelerate therapeutic decisions.

Fig. 1 .
Fig. 1.Comparison of PCT concentration between reference and study groups.Boxes show the median and interquartile range, whiskers show 10-90 th percentiles.Mann-Whitney test between PCT results in septic and non-septic group regardless of the day of sample collection

Fig. 2 .Fig. 3 .
Fig. 2. Comparison of sCD163 concentration between reference and study groups.Boxes show the median and interquartile range, whiskers show 10-90 th percentiles.Mann-Whitney test between sCD163 results in septic and non-septic group regardless of the day of sample collection

Table 1 .
Patients characteristics, divided into study and reference group according to the diagnosis of sepsis at the moment of sample collection

Table 2 .
PCT, sCD163 and NPT concentration in all patients, study and reference group

Table 3 .
ROC curve data for PCT, sCD163 and NPT

Table 4 .
Comparison of ROC curves