The Role of IL-6, 8, and 10, sTNFr, CRP, and Pancreatic Elastase in the Prediction of Systemic Complications in Patients with Acute Pancreatitis

Background and Aim. Early assessment of severity in acute pancreatitis (AP) is a key measure to provide rational and effective management. The aim of our study is to determine the prognostic value of interleukins (IL) 6, 8, and 10, soluble receptor for tumor necrosis factor (sTNFr), pancreatic elastase (E1), and C-reactive protein (CRP) as predictors of systemic complications in AP. Patients and Methods. A hundred and fifty patients with confirmed AP were enrolled in the study. The severity of AP was defined according to Atlanta criteria. Measurements of interleukins and sTNFr were performed on the first day of admission. CRP and E1 levels were assessed on admission and after 48 hours. ROC analysis was performed for all parameters. Results. Interleukins and sTNFr significantly differentiated patients with systemic complications from those without. Elevation of IL-6 showed the highest significance as a predictor (P = 0.001). CRP and elastase levels did not differ between mild and severe cases on admission, but reached statistical significance when measured on the third day (P = 0.002 and P = 0.001, resp.). Conclusion. Our study confirmed that IL-6, IL-8, IL-10, and sTNFr measured on admission, and CRP and pancreatic elastase measured on third day of admission represent valuable prognostic factors of severity and systemic complications of AP.


Background and Aim
Acute pancreatitis (AP) is a common and potentially lethal acute in�ammatory disease with an estimated overall mortality rate of 2% to 5%, and a signi�cant burden of morbidity and health care costs [1]. Although usually self-limiting, up to 20% of patients develop a severe form of disease, which can lead to a systemic in�ammatory response and multiple organ dysfunction and failure [2]. e two prevailing causes of AP are excessive alcohol consumption, most common in men, and gallstones, most common in women, which seem to act through different pathogenic mechanisms to induce pancreatic acinar cell damage. Several multifactorial scoring systems and routine clinical and biochemical parameters measured on admission and during the �rst 48 hours of hospitalization are used to estimate severity and promptly provide a rational and effective management.
Systemic manifestations of a disease initially limited to the pancreas are thought to be mediated by a variety of proand anti-in�ammatory mediators released from the pancreas and various other sources during the course of the disease. Several cytokines play a crucial role in the pathogenesis of AP by driving the additional in�ammatory response which leads to tissue damage and organ dysfunction. Local recruitment and activation of in�ammatory cells in AP may lead to the production of proin�ammatory cytokines, such as interleukins (IL) 6, 8, and tumor necrosis factor alpha (TNF-alpha) or his soluble receptor (sTNFr), as well as anti-in�ammatory IL-10. ese mediators have been mostly studied as markers of severity of acute pancreatitis. Another commonly applied and one of the �rst used markers for this purpose was Creactive protein (CRP). Different studies showed that a CRP value over 200 mg/L obtained at 48 hours aer onset of symptoms is highly predictive of pancreatic necrosis [3]. e use of pancreatic elastase for the differentiation between mild and severe cases of AP has also been investigated, yielding however con�icting results [4].
Still, insufficiently is known of the relationship between the clinical course of AP in humans and the dynamic of the major cytokines, in the presence or absence of pancreatic necrosis and distant organ complications. e purpose of our study was to determine the potential clinical value of interleukins (IL-6, IL-8, IL-10), soluble receptor for tumor necrosis factor (sTNFr), pancreatic elastase, and C-reactive protein as biochemical markers for predicting development of systemic complications in patients with AP.

Patients.
A hundred and �y patients with acute pancreatitis were prospectively entered into the study during a two-year period. e diagnosis of AP was made on the basis of a consistent clinical picture combined with a 3-fold increase of serum amylase or a 3-fold increase of serum lipase, and consistent morphological �ndings obtained by an ultrasound scan and/or computed tomography scan within the �rst 72 hours of admission. e severity of AP was assessed according to the Atlanta classi�cation. All patients irrespective of disease severity were included in the study. Mild acute pancreatitis (MAP) was de�ned as con�rmed AP without signs of major complications, while severe acute pancreatitis (SAP) was associated with the development of one or more local or systemic complications. Local complications included pancreatic tissue necrosis, as well as the formation of acute �uid collections, pancreatic pseudocyst, and abscess. Systemic complications assumed the presence of persistent systemic in�ammatory response syndrome (SIRS) and/or developing organ failure. SIRS was de�ned by 2 or more of the following criteria for >48 hours: heart rate >90 beats/min; rectal temperature <36 ∘ C or >38 ∘ C; white blood count <4000 or >12,000 per mm 3 ; and respirations >20/min or pCO 2 < 32 mmHg. Organ failure was de�ned as shock (systolic blood pressure <90 mmHg), pulmonary insufficiency (pO 2 < 60 mmHg), renal failure (creatinine >2 mg/dL, despite rehydration), and gastrointestinal bleeding (>500 mL/24 hours) [5,6]. e study was performed according to local ethics committee regulations.

Methods.
Blood samples from patients were obtained on admission and aer 48 hours. Samples for IL-6, IL-8, IL-10, IL-15, IL-17, and sTNFr were aliquoted in portions and stored at −20 ∘ C, not longer than two months. Measurements were performed using a commercially available ELISA kit (R&D Systems Inc., Minneapolis, USA) on a standard ELISA reader according to the manufacturer's instructions. Pancreatic elastase was analysed using a commercially available ELISA kit (ScheBo-Biotech, Giessen, Germany). Levels of CRP and other routine laboratory assessments were completed on biochemistry analyser Olympus AU 640 (Mishima Olympus, Japan). Measurements of interleukins and sTNFr were performed on samples obtained on the �rst day of admission. CRP and pancreatic elastase levels were assessed on admission and aer 48 hours.

Statistics.
All variables are expressed as medians with 95% co�dence intervals (95% CI). Mann Whitney test was used for comparison of independent samples. For differences between values of same parameters obtained on admission and aer 48 hours Wilcoxon pair test for dependent samples was used. Receiver operating characteristic (ROC) curves and respective areas under curve (AUC) were established for biochemical prognostic factors. Cut-off values were chosen as values that achieved the highest sensitivity and speci�city, as well as positive (PPV) and negative predictive values (NPVs). e proportion of patients without systemic complications was used as a measure of prevalence in performing ROC analysis. A value of < was considered statistically signi�cant.
e average value of IL-6 measured in the group of patients was higher than the upper limit of reference range recommended by the manufacturer (29 versus 12.5 pg/mL, resp.), whereas average values of other measured cytokines were within normal ranges. CRP measured on the �rst and third day of admission was above the upper limit of normal, as well as the average value of pancreatic elastase measured on the �rst day. Average values of pancreatic elastase measured on the third day were within the boundaries of recommended values (Table 1).
In the assessment of disease severity, average values of CRP and pancreatic elastase differed signi�cantly between the �rst and third day of hospitalization, with a signi�cant increase in CRP values and a signi�cant decrease in serum concentrations of pancreatic elastase (Table 2). e comparison of the analyzed biochemical prognostic factors between acute pancreatitis patients who developed systemic complications and those who did not is shown in Table 3. We found a signi�cant difference between the values of IL-6, IL-8, IL-10, and sTNFr evaluated on the �rst day of admission, and a signi�cant difference between CRP and elastase values analyzed from samples taken on the third day. No signi�cant difference was noted in the values of CRP and elastase on the �rst day between these two groups of patients.
e effectiveness of the investigated biochemical parameters in the early recognition of patients with and without systemic complications was assessed using ROC analysis. Considering the area under the ROC curve, values of cytokines measured on the �rst day were statistically significant indicators for development of systemic complications. CRP and pancreatic elastase measured on the third day also reached statistical signi�cance. Results from analysis are shown in Table 4. e largest area under the curve was for IL-6 (AUC = 0.71) and elastase on the third day (AUC = 0.70) ( Figures  1 and 2). Elastase had a fairly high sensitivity of 92%, but a rather low speci�city of 43%. e highest speci�city (84%) was calculated for the marginal value of CRP measured on the third day, with a sensitivity of 54% (Figure 3). CRP and elastase measured on the �rst day had the lowest predictive value (AUC = 0.51 and 0.56, resp.) not reaching statistical signi�cance.

Discussion
In this study we examined the value of IL-6, IL-8, IL-10, sTNFr, CRP, and pancreatic elastase as predictors of systemic complications in AP. e need for an early risk recognition and determination of best possible treatment modalities led to a series of investigations trying to establish an objective, rational, and clinically manageable severity assessment tool in patients with AP.
e initial acinar cell damage in the early stage of acute pancreatitis of any etiology is caused by a hypersecretion of pancreatic proteolytic enzymes. As a result there is an overproduction of in�ammatory mediators and free oxygen radicals. Tissue macrophages are the main source of proin�ammatory and anti-in�ammatory cytokines that attract neutrophils and more macrophages, and induce the production of proteases, elastases, and phospholipases. ese enzymes, as well as free oxygen radicals cause tissue damage, mainly vascular endothelial necrosis which leads to circulatory stasis. e increase of proin�ammatory and decrease of anti-in�ammatory cytokines are crucial factors in the We performed data analysis by dividing our patients into two groups. In one group we had patients who developed systemic complications ( ), and in the other those who had none ( ). e proportion of patients with systemic complications in our study correlates with the published data [7]. Our results show that the average value of IL-6 in patients with AP was above the upper limit of reference range recommended by the manufacturer, while average levels in controls were within normal ranges. ROC analysis was performed to evaluate the prognostic value of IL-6 to distinguish patients with systemic complications from those without, showing that patients with IL-6 concentrations greater than 37.9 pg/mL can be considered high risk in terms of developing systemic complications. We found a sensitivity of 82%, and a speci�city of 65%, with a PPV of 35%, and an NPV of 94%. In our previous study, results differed slightly with a sensitivity and speci�city of 68.7% and 69.9%, respectively, and PPV of 50%, and NPV of 83.6% [8]. ese differences probably derive from the quality of available tests. Pezzilli et al. showed in their paper an AUC of 0.91, a sensitivity of 100%, and speci�city of 83% [9]. ese values in combination with serum lipase levels achieved a diagnostic and prognostic accuracy of 94%. e most likely cause is a smaller number of patients and the    who suffered from post-ERCP pancreatitis have signi�cantly higher concentrations of these cytokines [11]. Using a cutoff level of 36 pg/mL they found that the sensitivity and speci�city for recognition of post-ERCP pancreatitis were 100% and 87%, respectively. e role of the proin�ammatory IL-8 in prediction of severity of acute pancreatitis seems less valuable than IL-6. Although it reached statistical signi�cance ( ) in the differentiation of mild and severe disease forms at a threshold value of 42.5 pg/mL, it achieved a modest sensitivity and speci�city of 68% and 67%, respectively. Same conclusions were obtained in previous studies by Pooran et al., and Berney et al. [12,13].
Interleukin 10 has an anti-in�ammatory role inhibiting the synthesis and release of other proin�ammatory cytokines and free oxygen radicals from macrophages and T-helper lymphocytes. On the other hand, it shows a positive effect on the proliferation and differentiation of B lymphocytes promoting the production of immunoglobulins. Our results show that a limit for IL-10 that can be said to separate milder from more severe forms of AP is 7.2 pg/mL, with a sensitivity of 75%, and a speci�city of 56%. Our previous study showed a lower sensitivity, and same speci�city for a higher cut-off value (37 pg/mL) [8]. ese differences are most probably related to the already mentioned different sensitivity of various commercial tests. A statistically signi�cant elevation of IL-10 in patients with severe AP was obtained by other authors as well [14][15][16]. However, two other studies found signi�cantly lower values of IL-10 in patients with severe AP [17,18]. Authors speculated that an impaired immune response to in�ammation could be a possible cause. It seems that a balance between pro-and anti-in�ammatory cytokines is the key process in the course of AP and development of systemic complications. A reduced functional reserve of IL-10 and a higher IL-6/IL-10 ratio could lead to SAP and a worse prognosis. However, this is still a matter requiring further investigations.
Results of sTNFr analysis, as another proin�ammatory cytokine, were consistent with previously published results showing a signi�cant elevation of serum concentrations in patients with SAP [19].
e research included the analysis of two parameters that can be considered as valuable indicators of the course of disease� pancreatic elastase as a speci�c enzyme secreted by pancreatic acinar cells, and CRP as an acute phase protein, both increasingly produced and released in a state of acute in�ammation. Elevated concentrations of E1 were measured in both groups of patients on admission, but without significant difference between the groups. However, we noticed a signi�cant decline in concentration of E1 between the �rst and third day ( ), with a signi�cant difference in values between the two groups patients ( ). Our results show that patients with a value of E1 below a cutoff value of 1.5 ng/mL measured on third day of admission could be considered potentially at low risk of development of systemic complications.
A group of Australian authors in their examination of E1 concentrations in 29 patients with acute pancreatitis showed a sensitivity of 80%, and a sensitivity of 96% for a E1 cut-off value of 3.5 ng/mL measured on admission, with a sensitivity and speci�city of 100% and 96%, respectively, on day three. e authors stated concurring limitations of ELISA tests, including lack of sensitivity due to questionable quality of used antibodies, and problematic reference range given by the manufacturer [20]. Similar limitations are present in all studies measuring serum concentrations of small amounts of antigens. erefore, ELISA methods are not standardized and recommended for routine cytokine and elastase analyses, and should be used only for research purposes. Moreover, the implementation of the analysis in larger series, which includes sample collection and subsequent determination, makes these methods unsuitable for routine use. Development of homogenous immunochemical methods appropriate for automated and standardized determination of individual samples should increase the prognostic value and practical application of different biochemical parameters.
As well as for E1 analysis, the concentrations of CRP differed signi�cantly between day one and three ( ). Patients who developed systemic complications showed sig-ni�cantly higher levels of CRP ( ). ROC analysis showed and AUC of 0.69 for a cut-off value of 214 mg/L, with a rather low sensitivity and speci�city, 54% and 84%, respectively. Gürleyik et al. showed similar results with sensitivity of 85%, and speci�city of 74%, although for a cutoff value that was signi�cantly different than in our study, 193 mg/L versus 214 mg/L, respectively [21]. Other authors also point out the value of CRP as a tool for AP severity assessment [22][23][24].
We con�rmed that CRP and elastase analyzed on the third day of admission, in addition to the evaluation of IL-6, IL-8, IL-10, and sTNFr on the �rst day, represent a valuable diagnostic tool in the assessment of severity and course of disease in patients with acute pancreatitis. Nevertheless, CRP is still the only recommended and standardized method for a fast and relatively inexpensive determination of severity of AP. Routine use of proin�ammatory cytokines as predicting factors of severity of acute pancreatitis is still not feasible in most hospitals, due to high costs and inaccessibility of analytic methods. erefore, development of new and more accessible laboratory equipment, as well as methods of analysis could help the clinicians in the early recognition of development of systemic complications and improve the management of severe acute pancreatitis.