Correlation between Serological Biomarkers and Disease Activity in Patients with Inflammatory Bowel Disease

Background Current biomarkers have been routinely used noninvasive methods for assessing disease activity of inflammatory bowel disease (IBD), but none of them are specific. This study was aimed to determine the performance of the serological biomarkers for detecting disease activity in patients with IBD. Methods A prospective study that included 73 ulcerative disease (UC) subjects, 141 Crohn's disease (CD) subjects, and 30 of them complicated with C. difficile infection (CDI) were diagnosed at a single-institution IBD center. Disease activity was assessed using by Truelove and Witts criteria for UC and Harvey Bradshaw Simple Index for CD. Serological inflammatory biomarkers were compared in different severity groups. Receiver operator curve analyses assessed the performance of each biomarker in discriminating disease states. Results For UC patients, elevated monocyte counts, C-reactive protein (CRP), and decreased lymphocyte counts and lymphocyte/monocyte ratio (LMR) significantly differed between subjects with active and inactive UC. LMR of 3.1 was 76% sensitive and had a specificity of 67% for active UC. For CD patients, higher values of neutrophils, monocytes, neutrophil/lymphocyte ratio, CRP, fibrinogen, and lower values of LMR and hemoglobin were significantly different between subjects with active and inactive CD. None of the biomarkers included had a good correlation with disease activity (area under the ROC Curve < 0.70). Conclusions A low LMR represents an inexpensive, readily available test with a promising value to identify disease activity in UC patients, whereas none of the inflammatory biomarkers showed a discriminative value in disease activity of CD.


Introduction
Inflammatory bowel disease (IBD), with its high incidence and prevalence, is now regarded as a worldwide healthcare issue [1][2][3]. Crohn's disease (CD) and ulcerative colitis (UC), the two major types of IBD, are chronic relapsing immunologic disorders of the bowel, and they appear to result from dysregulation of the immune system [4]. The assessment of IBD disease activity helps to guide clinical decisions of subsequent therapy [5]. Biomarkers in IBD can aid in the monitoring of disease severity in clinical practice, including erythrocyte sedimentation rate (ESR) and C-reactive protein (CRP) [6][7][8]. However, given the suboptimal performance of currently available biomarkers, endoscopy with biopsies remains the gold standard for evaluating and monitoring the inflammatory activity, but limited in use owing to its invasiveness and the need for on-demand specimen collection [9]. Thus, the search for easily accessible and cost-effective biomarkers that can be used to assess the disease activity is necessary and urgent for optimal management of IBD.
Recently, various serum markers, including the leukocyte differentials and neutrophil/lymphocyte ratio (NLR) have been evaluated as surrogate markers for predictive and prognostic values in various entities, such as rheumatoid arthritis, pancreatitis, and several malignancies [10][11][12][13]. Recent research focusing on cost-effective biomarkers in IBD has reinvigorated the examination of white blood cell patterns, and a few studies have showed the potential of neutrophil and lymphocyte counts and their ratio to assess the disease activity either in UC or CD [14,15]. Blood mononuclear cells were also found to correlate with UC activity [16]. However, the heterogeneity of the study populations led to variation, and there is a paucity of data regarding the utilities of other cell types and ratios reported, especially in CD patients. In this study, we aimed to elucidate the association of leukocyte values and the ratios with IBD activity and also differentiate it from Clostridium difficile infection (CDI).

Study Subjects.
The prospective study sample comprised 214 initially diagnosed IBD patients at the IBD unit from a single hospital between March 2017 and February 2018, including 73 UC and 141 CD patients, 30 of them with C. difficile infection (CDI). The diagnosis of UC and CD was separately based on standard clinical, radiological, endoscopic, and histological criteria. Informed consent was granted from all the patients, and ethical committee approval was obtained from the Ethics Committee of the Sir Run Run Shaw Hospital, Zhejiang University.
Data including patients' age at diagnosis, gender, and localization of the disease were retrieved from the medical database. Information of the patients' laboratory studies, endoscopic reports, and pathology reports at diagnosis was also recorded. The ratios of leukocyte fractions including neutrophil-to-lymphocyte ratio (NLR), lymphocyte-tomonocyte ratio (LMR), and neutrophil-to-monocyte ratio (NMR) were calculated. Patients were stratified according to the disease duration.
Exclusion criteria included (1) prior treatment with corticosteroids, hematological, neoplastic disorders, intestinal tuberculosis, or coagulopathy, (2) clinical evidence of active infection (except CDI), such as positive urine and blood cultures, documented skin infection, infiltrates on chest x-ray, and so on, and (3) patients <15 years of age.

Disease
Activity. The UC disease activity was assessed by Truelove and Witts criteria [5,17], as the criteria allow a simple and rapid stratification of UC patients and have been validated for over 60 years. To evaluate the UC disease activity index, the Mayo scoring system was used as previously described [18]. Patients with UC were classified as mild, moderate, or severe based on the number bloody stools per day, body temperature, pulse, hemoglobin, and ESR. Active UC was considered as having moderate or severe disease, whereas the remission period was defined as the mild group. For CD, the disease activity was classified by the Harvey Bradshaw Simple Index [19], based on the five variables (general well-being, severity of abdominal pain, number of liquid stools daily, presence of abdominal mass, and complications). Patients categorized as having an HBI score >4 were accepted as having active CD, whereas patients having an HBI score ≤4 were considered to be in the remission group.

Statistical Analysis.
Continuous variables with normal distribution were presented as mean ± standard deviation or, in the case of nonnormally distributed data, as median (range). Comparisons of proportion between different groups of patients were analyzed using the Chi-squared test. All normally-distributed values were carried out using student's t-test. Multivariate logistic regression analysis was used to evaluate the ratios (OR) and 95% confidence intervals (CI) for disease activity with significant findings in univariate analysis. Spearman's rank correlation coefficient was used to assess the correlations between the disease activity index of the disease and laboratory parameters (NLR, LMR, and NMR). Receiver operating characteristic (ROC) curve analysis was used to assess the accuracy of each biomarker (area under the curve, AUC) and identify optimal cutoff values of NLR, LMR, and NMR with maximum sensitivity and specificity for differentiation of activation of UC or CD from remission. A P value of less than 0.05 was considered statistically significant. The Statistical Package for Social Sciences (SPSS) 22 (Chicago, Illinois, USA) was used to analyze the data.

Results
A total of 214 patients diagnosed with IBD were included, 73 (34.1%) with UC and 141 (65.9%) with CD. One-hundred and forty-six (68.2%) were male, and the mean age of the whole cohort was 36 years (range 15-73 years). The demographic and disease characteristics of the patients included in the study are summarized in Table 1.

Diagnostic Biomarker Performance.
We identified several leukocyte ratio markers that were able to differentiate active UC/CD from UC/CD in the remission period, respectively. ROC analyses revealed that LMR (AUC � 0.722, 95% CI, 0.580-0.863) ratio was the best biomarker to differentiate active UC from UC remission patients. An LMR cutoff value of 3.1 had a sensitivity of 76% and specificity of 67% with lower values becoming progressively more specific (Figure 2). The diagnostic accuracy of the serological  biomarkers in CD patients was bad, as none of them had an AUC >0.7 (Figure 3).

Discussion
Recent studies have confirmed fecal calprotectin is the best biomarker for evaluating disease activity in IBD patients [20,21]. However, it is limited in clinical practice owing to its cost and discommodious sample collecting and processing.
In the present study, we evaluated the diagnostic accuracy of serological biomarkers to determine the disease activity in UC and CD patients, as the biomarkers were universally monitored at routine clinical practice and thus easily accessible. CRP and ESR are the most routine-used inflammatory indices for determining disease activity in patients with IBD [14]. However, the results of previous studies were disappointing, owing to the two biomarkers (CRP and ESR) with low sensitivity and specificity for reflecting the bowel inflammation [22,23]. Previous studies have shown   hypercoagulable state was associated with intestinal inflammation state, and serum fibrinogen level was correlated with the severity of the acute phase response [24,25]. A decreased serum albumin level has also been described to associate with increased systemic inflammatory load [26]. In this respect, in addition to the serum white blood differentials, we also investigated the biomarkers mentioned above.
In the current study, we found a significant association between elevated monocytes and CRP in patients with active UC compared with those with inactive UC, as well as decreased lymphocytes and LMR in patients with active UC, while further multivariate analysis showed a significant lower LMR was found in patients with active UC. Although there were more biomarkers, including the neutrophils, monocytes, HB, CRP, ESR, fibrinogen, NLR, and LMR, which were described associated with the disease activity of CD, only higher neutrophils and CRP were found in patients with active CD if assessed by multivariate analysis. The diagnostic accuracy of the serological biomarkers in CD patients was disappointing, as none of them had an AUC >0.7.
To get the more accurate results, we excluded the patients who had got medications (corticosteroid, thiopurines, etc) before the study, taking into account the potential influences that medications could have on the outcomes. Besides, infections are also confounders that can affect the leukocyte differential counts. Therefore, CDI, one of the active infections, was also examined in this study. We found elevated CRP and fibrinogen levels were associated with UC patients with CDI, comparing with inactive UC patients. Also, CRP levels in CD patients with CDI were statistically significant and higher than inactive CD patients.
The neutrophils play a key role in the active inflammatory response and are proposed to contribute to the destructive tissue cascades by secretion of interleukin-1, interleukin-6, and tumor necrosis factor-α. Moreover, previous studies in patients with IBD have strongly revealed that their lymphocyte function is abnormal at both the peripheral and mucosal level [27]. NLR was first identified as a parameter of systemic inflammation in 2001 [28], and it has been extensively reported thereafter in both malignant and nonmalignant conditions. Previous studies described NLR was a controversial marker in IBD. Elevated NLR was found in patients with active UC in comparison with healthy controls [14,29]. However, two recent studies showed NLR was effective in differentiating active UC/CD from healthy controls, but not from inactive UC/CD [15,16]. Through this study, we also did not find the diagnostic value of NLR in the disease activity of either UC or CD.
Factors such as medications can influence the leukocytetype counts. We know steroids can increase neutrophil count and subsequently the NLR. In this study, as mentioned above, we accounted for the potential confounders by enrolling only the initially diagnosed IBD patients, and other leukocyte subtypes and ratios were examined. Monocytes, a subset of leukocytes, differentiate into macrophages and dendritic cells in the inflamed tissues, involving in innate immunity by releasing proinflammatory cytokines, chemokines, and pathogen-associated molecular patterns [30][31][32]. Thus, activation of monocytes is prospected to be initiated during the active phase of IBD. Cherfane et al. performed a retrospective study in UC patients and reported the monocytes and LMR were promising biomarkers in UC. Similarly, our data revealed a significant association between monocytosis and disease activity of IBD, as well as the LMR. Besides, LMR had the best AUC (0.722) in UC patients. An LMR value lower than 3.1 carries a 76% sensitivity value for active UC. However, its diagnostic accuracy in differentiating disease activity of CD was undesirable. Our findings can be explained by the role that monocytes and lymphocytes play in the innate immune responses in such an inflammatory disease as IBD.
There are several strengths for this study. Firstly, it is a prospective study to prove the utilities of inflammatory biomarkers for severity stratification in patients with UC and CD. To our knowledge, there are limited data for analyzing the efficacy of these biomarkers in CD. Secondly, the study cohort is homogenous, in which the diagnosis and severity disease evaluation were allocated based on standardized definitions. Additional strengths of our study were inclusion of active disease, quiescent disease, CDI, and taking into account the influence of medications. However, this was a single-center cohort with a relatively small sample size.
In conclusion, among the inflammatory biomarkers, including CRP, ESR, fibrinogen, leukocytes, NLR, LMR, and NMR, the LMR has the highest discriminatory capacity for severe UC, with an optimal cutoff value of 3.1, but none of them with a discriminative value in evaluating disease activity of CD. Further work with multicenter studies to assess the biomarkers is warranted.

Data Availability
The excel data used to support the findings of this study are available from the corresponding author upon request.

Ethical Approval
This study was conducted with the approval of the Ethics Committee of Sir Run Run Shaw Hospital, School of Medicine, Zhejiang University.