Potential blood biomarkers in chronic spontaneous urticaria

Chronic spontaneous urticaria (CSU) is a mast cell‐driven disease that is defined as the recurrence of weals, angioedema or both for > 6 weeks due to known or unknown causes. As of yet, disease diagnosis is purely clinical. Objective tools are needed to monitor the activity of CSU and the efficacy of treatment. Recently, several reports have suggested that blood parameters may be considered as potential disease‐related biomarkers. Here, we reviewed available literature on blood biomarkers for CSU diagnosis, activity monitoring, duration, patient subgroup allocation or response to treatment. We performed a PubMed, Google Scholar and Web of Science search and identified and analysed 151 reports published prior to January 2016. We found strong evidence for significant differences between patients with CSU and healthy controls in blood levels or values of D‐dimer, C‐reactive protein (CRP), matrix metalloproteinase‐9 (MMP‐9), mean platelet volume (MPV), factor VIIa, prothrombin fragment 1 + 2 (F1 + 2), tumour necrosis factor, dehydroepiandrosterone sulphate and vitamin D. Also, there is strong evidence for a significant association between CSU activity and blood levels or values of D‐dimer, F1 + 2, CRP, IL‐6 and MPV. Strong evidence for reduced basophil count and high levels of IgG anti‐FcεRI in the subgroup of CSU patients with positive autologous serum skin test was shown. In contrast, the evidence for all reported blood biomarkers for differentiating CSU from other diseases, or a role in prognosis, is weak, inconsistent or non‐existent. Taken together, we identified 10 biomarkers that are supported by strong evidence for distinguishing patients with CSU from healthy controls, or for measuring CSU activity. There is a need for further research to identify biomarkers that predict outcome or treatment response in CSU.


Introduction
Chronic spontaneous urticaria (CSU) is a mast cell-driven disease that is defined as the recurrence of weals, angioedema (AE) or both for > 6 weeks due to known or unknown causes [1]. At any time, 0.5-1% of the population suffer from CSU [2]. The disease has a large impact on the quality of life [3], and more than 30% of patients with moderate to severe symptoms still have CSU after 5 years [4].
Assessment of disease activity, its impact on quality of life, treatment efficacy and disease control, at present, is purely clinical and includes patient medical history, physical examination and the use of questionnaires such as the Urticaria Activity Score (UAS), the Angioedema Activity Score (AAS), the Chronic Urticaria Quality of Life (CU-Q2oL), the Angioedema Quality of Life (AE-QoL) and the Urticaria Control Test (UCT) [5]. All of these tools are validated and of great value in the monitoring of patients with chronic urticaria, but all have the disadvantage of being subjective, retrospective instruments. Thus, there is a lack of objective measurable indicators of the activity of the disease, and the efficacy of treatment.
Recently, several reports have suggested that blood parameters may indicate disease activity and duration and help with monitoring treatment and may be considered as potential prognostic biomarkers of CSU [6][7][8][9]. According to the National Institute of Health (NIH) Biomarkers Definitions Working Group, a biomarker is a 'characteristic that is objectively measured and evaluated as an indicator of normal biological processes, pathogenic processes or pharmacological responses to a therapeutic intervention' [10]. Important characteristics of a biomarker are its specificity, sensitivity and validity, and ideally its correlation with the activity and prognosis of the disease. The discovery, validation and implementation of a reliable biomarker could eventually lead to a personalized approach to the assessment and prediction of CSU course and the treatment for the disease and would be of great help as an objective follow-up for patients in everyday life as well as in clinical trials.
We performed a review to clarify the blood parameters that could be possible biomarkers of CSU differential diagnosis, activity, duration, patient subgroup allocation and response to treatment. Inclusion criteria were as follows: (1) the studies that include patients with CSU. In our analyses, we excluded cases of acute and inducible urticaria whenever possible. However, we did not exclude studies in which no distinct differentiation between CSU and chronic inducible urticaria had been made, but included this information in the tables' legends; (2) the reports in which laboratory blood parameters (protein levels in serum or plasma) including cell count were or were not significantly associated or correlated with one of the following outcomes: CSU differential diagnosis, activity, duration, patient subgroup allocation and response to treatment.

Methods
The following key exclusion criteria were used in this review: (1) non-relevant studies, for example studies that did not include patients with chronic urticaria, where blood markers were not evaluated or an association between laboratory and clinical parameters was not analysed; (2) due to space limitations and a high number of studies available, we excluded other biomarkers, or potential biomarkers, such as skin protein expression, or functional tests including basophil activation test and autologous serum skin test (ASST). These should be evaluated separately. However, we assessed the levels of potential blood biomarkers in patients with positive and negative ASST results.
We identified 2185 studies; 2034 of which were excluded from further analysis after review of titles, abstracts or full texts: non-relevant studies (n = 1964), reviews (n = 31), case reports (n = 27) and articles without abstract or full text in English (n = 12). The remaining 151 studies were included in our analyses. They comprised 34 prospective studies, 11 retrospective studies, 109 case-control studies and 29 cross-sectional studies; 97 studies included only patients > 16 years, and 3 studies included only patients < 16 years.
The rating system developed by de Croon et al. [11] was applied to each potential biomarker (Table 1). We categorized the levels of scientific evidence as 'no evidence', 'inconsistent', 'weak' or 'strong' based on the outcome and number of studies from different centres and teams (e.g. if one team published three studies on the same biomarker that show the same association, we counted it as one study). It is important to show a consistency (reproducibility) of association between the studies; that is, the association is consistent when results are replicated in studies in different settings using different methods [12]. Many studies were relevant for more than one of outcomes and were therefore included in more than one table. Only statistically significant findings with a P value of < 0.05 were included. However, data were not adjusted for potential confounders.

Potential biomarkers for confirming CSU and distinguishing it from other diseases
One hundred and thirty-three studies reported the presence or absence of significant differences in blood Table 1. Definitions of levels of evidence [11] Description Levels of evidence for 'association' Strong Three studies available that find an association in the same direction or ≥ four studies available, of which > 66% find a significant association in the same direction and no more than 25% find an opposite association Weak Two studies available that find a significant association in the same direction or three studies available, of which two find a significant association in the same direction and the third study finds no significant association  (Table 2). Strong evidence in blood levels of biomarkers between patients with CSU and patients with other diseases was not demonstrated (Table S2). Weak evidence was shown for D-dimer (higher levels in CSU vs. psoriasis), total IgE (lower levels in CSU vs. acute urticaria) and IgG anti-FceRI (higher levels in CSU vs. atopic dermatitis).

Potential biomarkers of disease activity
Seventy-five studies reported a correlation between biomarkers and CSU activity (Tables 2 and 3). Strong evidence of a significant association between blood levels or values of biomarkers and CSU activity was demonstrated only for D-dimer, CRP, F1 + 2, MPV and IL-6. The evidence for a significant association between CSU activity and blood levels of IL-17, MMP-9 and FVIIa was weak.

Potential biomarkers for CSU duration
Fifteen studies reported association between biomarker blood levels and CSU duration (Table 4), with very mixed results. Weak evidence for positive antithyroid antibodies and longer CSU duration was demonstrated. The evidence for other biomarkers is inconsistent or non-existent.

Potential biomarkers for CSU subgroup allocation
Seventy-three studies reported differences in blood levels of biomarkers between subgroups of patients with CSU ( Table 5, Table S3). Strong evidence for reduced basophil count and high levels of IgG anti-FceRI in the subgroup of CSU patients with positive autologous serum skin test was shown. Weak evidence for higher MPV and IL-17 in ASST-positive CSU patients was demonstrated. The evidence for other biomarkers and the presence of angioedema, age, gender, atopy, ASST and autologous plasma skin test positivity is inconsistent or non-existent.

Potential biomarkers of response to treatment
Sixteen studies reported a correlation between levels of biomarkers and response to treatment or an association between levels of biomarkers before and after treatment ( Table 6). Levels of tryptase, D-dimer and IL-31 were evaluated separately in three different studies during the course of omalizumab treatment. Significant association between levels of biomarkers before and after treatment was shown for D-dimer and IL-31, but not for tryptase. Levels of these parameters were elevated before treatment with consequent decrease after therapy with omalizumab. Low levels of IL-31 were associated with response to treatment. Levels of D-dimer, CRP and IL-2R, IL-5 and TNF were separately evaluated in three studies during the course of ciclosporin treatment. Here, decreased D-dimer levels were significantly associated with response to treatment. Levels of CRP and IL-2R, IL-5, TNF were elevated before treatment with consequent decrease after therapy with ciclosporin.
There have been no studies conducted on biomarkers that predict response to treatment.
Thus, the level of evidence for all biomarkers is nonexistent.

Discussion
In recent years, many publications have reported about potential biomarkers in CSU. Here, we performed a review to obtain an overview about the potential of these reported biomarkers to be used in clinical practice. The six questions asked to the literature, therefore, were as follows: Are there potential biomarkers for the distinction of CSU from other diseases and healthy individuals? Are there potential biomarkers that correlate with CSU activity? Are there potential biomarkers that can predict CSU duration? Are there potential biomarkers that allow for subgroup allocation? Are there potential biomarkers that reflect response to treatment or predict the response? Is there a correlation between different potential biomarkers found in patients with CSU? A good biomarker should be valid in any or at least in most CSU patient populations. There is a wide variability and heterogeneity in CSU patient population from different teams, different centres and different countries (Tables S1 and S2; Tables 2 and 3). This heterogeneity of the data is the main reason why we could not conduct a meta-analysis.
The most promising results come from studies looking at D-dimer, CRP and VD, raising the question of their possible role in the pathophysiology of urticaria. From the available literature, we found that levels of D-dimer and CRP were significantly higher in patients with CSU in comparison with HCs in nine and eight studies from different teams, respectively. Levels of VD were significantly lower in patients with CSU in comparison with HCs in eight studies from different teams. D-Dimer is a fibrin degradation product that reflects the expression of tissue factor by eosinophils, the activation of the coagulation cascade and thrombin generation. The latter can increase vascular permeability and induce degranulation of mast cells [13]. CRP is a sensitive marker that may be increased in CSU due to mast cell activation accompanied by inflammatory response [14]. VD is thought to have potential immunomodulatory activity and now has been intensely studied for its role in various chronic diseases including CSU [15].
Less convincing data exist for MMP-9, MPV, FVIIa, F1 + 2, TNF, IgG anti-FceRI and DHEA-S. There have been conflicting reports on various other biomarkers such as SP, basophil count, total IgE, VEGF or IL-18 [16][17][18][19][20][21]. For example, we have described SP to be increased in patients with CSU and to correlate with disease activity [22], while Tedeschi and co-workers have reported no elevation in SP in comparison with normal subjects [23].
Some studies were carried out in only one centre and included small samples of patients with CSU that could be responsible for under-or overestimation of the results. For example, Confino-Cohen et al. [24] retrospectively described significantly higher values of MPV in 12 778 patients with CSU in comparison with 10 714 control subjects, whereas Isiksacan et al. [25] found significantly decreased values of MPV but in small sample of patients with CSU. Importantly, these data show that any valid biomarker may only be found by assessing large patient groups in multi-centre approaches.
Conflicting evidence might also be explained by different patient populations and differences in analysis of the results (e.g. difference in standardization -, not defined or no data; P, plasma; S, serum; MPV, mean platelet volume; PDW, platelet distribution width; PC, platelet count; CRP, C-reactive protein; dsDNA, double-stranded DNA; TPO, thyroid peroxidase; ATA, IgG antithyroid antibodies; MMP-9, matrix metalloproteinase-9; ANA, antinuclear antibodies; R, retrospective study; P, prospective study; CC, case-control study; *between levels or values of potential biomarkers and longer CSU duration; † chronic inducible urticaria was not excluded or it is unclear from the paper. between the various producers of the assay kits, the existence of various cut-offs and distinct methodological measurements of blood parameters). For instance, CRP levels were higher in patients with CSU than in HCs as measured by immunoturbidimetric fully automated assay in one study [26] and were not significantly different between idiopathic CSU and HCs as measured by ELISA in another [27]. Another example would be poor agreement between manual and automated basophil counts in patients with CSU [21]. In addition, the inconsistent and controversial results found in the literature might be explained, at least in part, by the treatment of patients with drugs such as corticosteroids, and the lack of exclusion of concomitant diseases that could have influence on the levels of biomarkers. A general problem in the search for a potential biomarker is the lack of specificity of the respective    *dyspeptic controls; † ECP was evaluated at enrolment and 8 weeks after eradication; -, not defined or no data; P, plasma; S, serum; VD, vitamin D; AH, sedative antihistamines; nsAH, nonsedating antihistamines; CS, corticosteroids; SCF, stem cell factor; NPY, neuropeptide Y; NGF, nerve growth factor; VIP, vasoactive intestinal peptide; CGRP, calcitonin gene-related peptide; SP, substance P; CRP, C-reactive protein; IL, interleukin; TNF, tumour necrosis factor alpha; DAO, diamine oxidase; ECP, eosinophil cationic protein; R, retrospective study; P, prospective study; CC, case-control study.
marker. For example, D-dimer or CRP levels have been shown to be elevated in CSU and correlate with disease activity. However, the problem remains as to whether these substances are specific enough for urticaria, as they can be elevated in many other diseases that are often comorbid in patients with CSU or even an underlying cause of CSU (e.g. chronic infections, autoimmune disorders).
There are other factors that should also be taken into consideration: differences in the design of studies and different age of patients between studies, healthy controls and patients with CSU were not perfectly matched in some studies, and in some cases, confounding factors could have had an influence on the outcome of potential biomarker assessment (e.g. levels of VD depend on the season and diet).
Type I autoimmunity and type II autoimmunity are thought to be involved in the subset of patients with CSU as demonstrated by the presence of IgG and IgE autoantibodies [28][29][30]. We evaluated the results of available studies where immunoassays (Western blot, ELISA, radioimmunoassay or immunoenzymatic assay) were used for the detection of autoantibodies. There is inconsistent and weak evidence that patients with CSU have higher levels of IgG anti-FceRI than HCs and patients with atopic dermatitis, respectively.
The conflicting results of studies and the presence of autoantibodies in HCs and other diseases may be explained by different methodology, the concept of 'conditional autoimmunity' and non-functional autoantibodies. For example, the direct ELISA approach used in some studies can lead to masking of IgE antibodies due to competition with IgG antibodies [30,31]. 'Conditional autoimmunity' suggests that autoantibodies can cause CSU only under certain conditions and they become pathogenic over time [32]. Fiebiger and coworkers described functional histamine-releasing IgG anti-FceRI antibodies in patients with CSU and nonfunctional and non-pathogenic IgG anti-FceRI in patients with systemic lupus erythematosus, pemphigus vulgaris, dermatomyositis and bullous pemphigoid [33].
The minimum requirement for a disease biomarker is a correlation with disease activity. The most promising candidates so far are D-dimer, CRP, F1 + 2, IL-6 and MPV, whose levels or values were significantly higher in more severe CSU in 10, 7, 5, 4 and 3 studies from different investigators, respectively. Elevated plasma levels of F1 + 2 (much like D-dimer) suggest activation of coagulation with thrombin generation [34]. MPV is a potential marker of platelet reactivity and may reflect the ongoing inflammation [24].
Inconsistent or weak evidence on other parameters such as total IgE, FVIIa, MMP-9, IL-17, IL-18, DHEA-S, SP, prolactin, VEGF, TNF and IgG anti-FceRI may be associated with various symptom scoring systems used to assess CSU activity, sample size, age of patients, number of centres, serum or plasma used and design of study (Table 3). UAS or UAS7 was applied in 37 studies. In some studies, only patients with active urticaria or with positive ASST were recruited. Another concern was the different methodology among the studies. For example, Kessel et al. [17] showed that high levels of total IgE were significantly associated with more severe CSU as measured by ELISA with a cut-off of 175 U/mL. In contrast, Baek et al. [35] used immunoradiometric assay with a cut-off of 91 IU/mL and found no association between total IgE and CSU severity.
There is weak or conflicting evidence on the association between levels of antithyroid antibodies, VD, total IgE and disease duration (Table 4). Such controversy may be, in part, explained by differences in patient selection, sample size and study design as well as in the duration of the follow-up period. For example, Woo et al. [15] did not analyse factors that affect the levels of serum VD (e.g. smoking, daily sunlight per subject, baseline nutritional status, time of blood sampling) due to retrospective design of study. Moreover, other investigations of VD were limited by small sample size, selection and detection bias [36][37][38], and in some studies, no specific criteria for urticaria remission were described [4,37,39]. For other parameters such as IgG or IgE autoantibodies, no significant association was found. Taken together, there is currently no known predictor of disease duration and further studies are required to identify whether such predictors do exist.
Up to 40-50% of patients with CSU have autoreactive urticaria as defined by positive ASST [40]. We found strong evidence for a high prevalence of IgG anti-FceRI in ASST-positive CSU patients [41][42][43]. Three studies from different centres and teams provided evidence that positive ASST is linked to reduced basophil numbers [44][45][46]. Evidence for basopenia in autoimmune CSU is consistent with the hypothesis that circulating basophils may be recruited from blood into urticarial weals during disease activity [21]. Significant associations between ASST positivity and levels of potential biomarkers were shown in some studies for MPV, IL-17, IgG antithyroid antibodies, CRP, IL-18, ANA, but not in other. IgE anti-TPO and anti-dsDNA were not associated with autoreactivity. Thus, positive ASST may be linked only to IgG autoantibodies or other histamine-releasing factors.
It is as of yet unclear why some patients with CSU present with weals only, while others have weals and angioedema or angioedema only. Therefore, it would be of great interest to identify biomarker associated with the absence or presence of angioedema in CSU. However, no such biomarker has been identified yet and the presence of angioedema was shown to be not associated with altered levels of anti-FceRI, anti-IgE, SP, B cell activation factor and tryptase. Therefore, future investigations are needed to reliably identify CSU subgroups.
The emergence of new therapies in the treatment for urticaria emphasizes the demand of a biomarker in order to objectively assess the response to therapy or to predict the outcome of a therapy. The constant need for new therapies aims at the maximum efficacy and the best fit for individual patients. Studies fall short on measuring the response to treatment, forgetting often to correlate changes in substances and UAS at different time-points. Significant association between levels of biomarkers before and after treatment with omalizumab, ciclosporin or vitamin D supplementation was shown for D-dimer, IL-31 (omalizumab), CRP, IL-2R, IL-5, TNF (ciclosporin) and VD, respectively. Significant association was shown between levels of D-dimer and IL-31 and response to treatment with ciclosporin and omalizumab, respectively. These two biomarkers may be the most promising candidates, but further research is needed to verify this and/or to identify other potential biomarkers.
A correlation between biomarkers may be important in understanding of pathogenetic processes in CSU and evaluation of two or more biomarkers in the same patient may increase a prognostic value.
Non-conflicting weak evidence was only shown for a positive correlation between the levels of D-dimer and CRP as well as CRP and MMP-9. The elevation of Ddimer and CRP in CSU and correlation between them and with disease activity may support the involvement and the close link between coagulation activation and inflammation in CSU pathogenesis [34].

Conclusion
We identified 10 biomarkers that are supported by strong evidence for distinguishing patients with CSU from healthy controls, or for measuring CSU activity. There is a need for further research to identify biomarkers that predict outcome or treatment response in CSU.

Funding sources
None declared.

Supporting Information
Additional Supporting Information may be found online in the supporting information tab for this article: Table S1. Are there potential biomarkers that allow for distinction of CSU and healthy controls? Table S2. Are there potential biomarkers that allow for the distinction of CSU from other diseases? Table S3. Are there potential biomarkers that allow for subgroup allocation? Table S4. Is there a correlation between different potential biomarkers found in CSU patients?