Efficacy and safety of antagonists for chemoattractant receptor-homologous molecule expressed on Th2 cells in adult patients with asthma: a meta-analysis and systematic review

Background Chemoattractant receptor-homologous molecule expressed on Th2 cells (CRTH2) antagonists are novel agents for asthma but with controversial efficacies in clinical trials. Therefore, we conducted a meta-analysis to determine the roles of CRTH2 antagonists in asthma. Methods We searched in major databases for RCTs comparing CRTH2 antagonists with placebo in asthma. Fixed- or random-effects model was performed to calculate mean differences (MD), risk ratio (RR) or risk difference (RD) and 95% confidence interval (CI). Results A total of 14 trails with 4671 participants were included in our final analysis. Instead of add-on treatment of CRTH2 antagonists to corticosteroids, CRTH2 antagonist monotherapy significantly improved pre-bronchodilator FEV1 (MD = 0.09, 95% CI 0.04 to 0.15, P = 0.0005), FEV1% predicted (MD = 3.65, 95% CI 1.15 to 6.14, P = 0.004), and AQLQ (MD = 0.25, 95% CI 0.09 to 0.41, P = 0.002), and reduced asthma exacerbations (RR = 0.45, 95% CI 0.23 to 0.85, P = 0.01). Rescue use of SABA was significantly decreased in both CRTH2 antagonist monotherapy (MD = − 0.04, 95% CI -0.05 to − 0.03, P < 0.00001) and as add-on to corticosteroids (MD = − 0.78, 95% CI -1.47 to − 0.09, P = 0.03). Adverse events were similar between the intervention and placebo groups. Conclusions CRTH2 antagonist monotherapy can safely improve lung function and quality of life, and reduce asthma exacerbations and SABA use in asthmatics. Electronic supplementary material The online version of this article (10.1186/s12931-018-0912-y) contains supplementary material, which is available to authorized users.


Introduction
Asthma is a common respiratory disease characterized by chronic airway inflammation, airway hyperresponsiveness, and reversible airflow limitation, which affects more than 300 million people worldwide and imposes a considerable social and economic burdens [1]. Most of patients can be effectively controlled by inhaled corticosteroids, the first-line therapy as recommended by the Global Initiative for Asthma (GINA) guideline [2], however, at least 40% of asthmatics remain inadequately controlled in spite of treatment with high dose of inhaled corticosteroids [3]. Moreover, a clear association between risk of adverse effects and long-term use of corticosteroid has also been observed [4], therefore, novel therapeutics is warranted to improve symptoms control and avoid overuse of steroids.
The chemoattractant receptor-homologous molecule expressed on Th2 cells (CRTH2) is a G-protein coupled receptor, and it is reported to be crucial in asthma development due to the chemotaxis of type 2 helper T cells and eosinophils, delay in cell apoptosis, as well as production of proinflammatory cytokines including interleukin-4, 5, and 13 by the activation of prostaglandin D 2 (PGD 2 ) [5][6][7]. Accumulating evidence has shown that the blockade of CRTH2 receptor significantly reduces allergic airway inflammation in animal models [8][9][10], but inconsistent efficacy and safety profiles of CRTH2 antagonists are noticed in clinical trials. Barnes and his colleagues [11] for the first time reported that OC000459, a CRTH2 antagonist, significantly improved quality of life but had no effect on lung function and airway inflammation in patients with asthma, while a significant improvement of forced volume in one second (FEV 1 ) [12] and inhibition of post-allergen increase in sputum eosinophils [13] but no relief of asthma symptoms in symptomatic controller-naïve asthmatics [14] were reported by subsequent studies.
Based on the current controversial and ambiguous findings in the treatment of patients with asthma by CRTH2 antagonists, we conducted a meta-analysis and systematic review of all available randomized controlled trials (RCTs) to further determine the roles of CRTH2 antagonists in asthmatics.

Search strategies
A comprehensive computer search was conducted in Cochrane Central Register of Controlled Trials (CEN-TRAL), Pubmed, Medline, Embase, ISI Web of Science and American College of Physician (ACP) between 1946 and September 2018 by using the keywords of "CRTH2" or "chemoattractant receptor-homologous molecule expressed on TH2 cells" or "chemoattractant receptor expressed on TH2 cells" or "DP2" or "prostaglandin D2 receptor" and "antagonist" or "inhibitor" and "asthma". Publication type and species were limited to RCTs and humans, respectively, but we did not limit the publication language. References listed in each identified article were checked and the related articles were searched manually to identify all eligible studies and minimize the potential publication bias.

Inclusion and exclusion criteria
Eligible clinical trials were identified based on the following criteria: 1) asthma was diagnosed by physicians according to the GINA guideline [2] with the evidence of airway hyperresponsiveness (the provocation concentration of methacholine causing a 20% fall in FEV 1 (methacholine PC 20 ) < 16 mg/mL) and/or bronchodilator responsiveness (an increase of FEV 1 % predicted > 12% and FEV 1 > 200 mL following inhalation of 200 μg salbutamol); 2) age was not less than 18 and smoking history was no more than l0 pack-years; 3) study designs were randomized placebo-controlled trials; 4) intervention treatment was oral CRTH2 antagonists regardless of dose, frequency, and durations; 5) outcomes included but not limited to lung function, asthma control and quality of life scores, sputum and blood eosinophil count, fractional exhaled nitric oxide (FeNO), asthma exacerbations, rescue use of short-acting β 2 agonists (SABA), and adverse events. Retrospective, observational, cohort or case control studies were excluded.

Study selection
Two investigators independently performed the study selection in two phases. First, they screened the titles and abstracts of all identified studies to discard duplicated and nonrandomized controlled studies. Then, eligible studies were extracted by reviewing full texts according to the previously defined study inclusion and exclusion criteria. Disagreements were resolved by consensus or consulting a third investigator.

Data extraction and quality assessment
Two investigators independently and separately conducted the data extraction and quality assessment. Data from eligible studies were extracted in a standard form recommended by Cochrane [15] including authors, publication year, study design, participant characteristics, population, interventions, concomitant treatment, outcome measures and study results. Cochrane risk of bias tool was used to assess the risk of bias in estimating the study outcomes. Each study was assessed for: 1) random sequence generation; 2) allocation concealment; 3) blinding of participants and personnel; 4) blinding of related outcomes assessment; 5) incomplete outcome data; 6) selective reporting; and 7) other biases. For any missing data or information, we contacted corresponding authors by e-mail to request the full original data. Any divergence was resolved by mutual consensus in the presence of a third investigator.

Statistical analysis
Statistical analysis was accomplished by an independent statistician using Cochrane systematic review software Review Manager (RevMan; Version 5.3.5, the Cochrane Collaboration) and Stata (version 14.0, Stata Corporation, USA). P value < 0.05 was defined as statistical significance and the results were showed in forest plots. We conducted a systematic review when data could not be pooled in meta-analysis.
Continuous variables were expressed as mean and standard deviation (SD), while dichotomous variables were shown as frequency and proportion. Mean differences (MD) and 95% confidence interval (CI) were calculated for continuous data, and risk ratio (RR) or risk difference (RD) combined with 95% CI for dichotomous data. If a study presents more than two interventions, they were combined into a single intervention group according to the Cochrane handbook [15]. Heterogeneity was quantified by I 2 statistic and chi-squared test with P < 0.1 and I 2 > 50% indicating significant heterogeneity. Random-effects model was applied in the statistical heterogeneity; otherwise fixed-effects model was used. Publication bias was tested by Funnel plot with Egger's and Begg's tests. All analyses were conducted based on the intention-to-treat principle. The potential influence of pre-specified factors, such as types of CRTH2 antagonists, presence of concomitant treatment, treatment duration, asthma severity, on the effect estimates was further explored via random-effects model meta-regression when an outcome of interest was reported by at least three RCTs in each subgroup.

Results
A total of 659 potentially relevant articles were identified, and 490 articles were screened for eligibility after removal of 169 duplicate records. After reviewing the titles and abstracts, we identified and retrieved 34 studies for later full-text assessment due to the discard of non-RCTs (n = 253), animal experiments (n = 90), non-CRTH2 antagonists (n = 87), non-asthmatic patients (n = 18), and others going against our inclusion criteria (n = 8). Finally, 14 studies were included for our systematic review and meta-analysis because 20 studies were excluded owing to abstract form of included studies (n = 12) and insufficient data for analysis (n = 8). (Fig. 1).

Quality assessment
Based on the six domains, all the included studies showed low risk of bias (Fig. 2). The method used in randomization sequence generation and allocation concealment was clearly described in all the studies except seven studies [13,[17][18][19][22][23][24]. All the 13 studies were double-blinded and reported complete outcome data.
No statistical heterogeneities were found except for that in morning PEF in the pooled analysis (I 2 = 55%, P = 0.08) (Fig. 4).

Asthma exacerbations
Six studies [11,12,16,17,22,23] presented data on asthma exacerbations, and they all included patients with exacerbations based on a decline of more than 30% from the baseline in morning PEF on two or more consecutive mornings, or a worsening of asthma symptoms requiring treatment with systemic corticosteroids or increased doses of rescue medication, and/or the need for asthma-related hospitalization/emergency room visit.

Sputum and blood eosinophils
Five studies [11,13,17,20,21] presented data on sputum eosinophils in patients with CRTH2 antagonists treatment, and three studies [20,21,24] reported blood eosinophils. However, we could not conduct individual synthesized analysis of each outcome due to the inconsistently reported data, in which eosinophils levels were presented either as amount per gram or percent of the whole white cells, and we were unable to extract mean change of eosinophils after treatment from only mean (range) or geometric mean (95% CI). Table 3 summarized available studies with sputum or blood eosinophils. Ambiguous results were noticed in sputum eosinophils as some studies [13,20] showed significant reduction of sputum eosinophils in patients with CRTH2 antagonists treatment compared to placebo while some studies [11,17,21] did not find any significant differences. As for blood eosinophils, all three available studies showed that CRTH2 antagonists treatment could not significantly reduce blood eosinophils compared to placebo regardless of being as monotherapy or an add-on treatment to corticosteroids.

FeNO and methacholine PC 20
Similar to sputum and blood eosinophils, data of FeNO and methacholine PC 20 from available studies could not   be pooled in meta-analysis. In total, six studies [13,16,17,20,21,24] and two studies [13,24] depicted change of FeNO and methacholine PC 20 , respectively. No significant difference in FeNO was found between CRTH2 antagonists and placebo even in CRTH2 antagonists monotherapy or as add-on treatment to corticosteroids. However, Diamant et al. [24] showed that Setipiprant monotherapy significantly stablized methacholine PC 20 in stable allergic steroid-free asthma compared to placebo, which was not observed in the study by Singh et al. [13] in stable allergic steroid-naïve asthmatics treated with OC000459 monotherapy.

Discussion
In our study, we found that CRTH2 antagonists, compared to placebo, significantly improved pre-bronchodilator FEV 1 (L) and AQLQ scores, reduced ACQ scores and SABA use in adults with asthma, which was also true in the treatment of CRTH2 antagonists monotherapy except for no effect on ACQ scores but improved pre-bronchodilator FEV 1 % predicted and reduced asthma exacerbations. However, CRTH2 antagonists as add-on treatment to corticosteroids did not show any obvious superior advantages over placebo. CRTH2 antagonists monotherapy was associated with lesser adverse events leading to treatment withdrawal, but CRTH2 antagonists, regardless of monotherapy or as add-on treatment to corticosteroids, showed similar incidence of adverse events, severe adverse events, and treatment related adverse events compared with placebo.
Reversible airflow limitation and airway hyperresponsiveness are the key traits in asthma pathophysiology, and FEV 1 , PEF as well as Methacholine PC 20 are the most widely used parameters to assess asthma severity and control, and predict future risk of asthma exacerbations [26,27]. Our study found that CRTH2 antagonists monotherapy could significantly improve pre-bronchodilator FEV 1 and FEV 1 % predicted, which might be attributed to the potential anti-inflammation effects of CRTH2 antagonists [5][6][7]. As  The effect of CRTH2 antagonists vs placebo on AQLQ. AQLQ, asthma quality of life questionnaire; CI, confidential interval; CRTH2, chemoattractant receptor-homologous molecule expressed on Th2 cells; SD, standard deviation; vs, versus mentioned above, the binding of PGD2 to CRTH2 induces respiratory burst and degranulation of eosinophils as well as increases release of type 2 cytokines, leukotrienes and cationic proteins, which may damage airway epithelia, thus resulting in airway narrowing and development of airway hyperresponsiveness [7,[28][29][30]. Furthermore, it has been demonstrated that production of type 2 cytokines is associated with greater decline in lung functions [31]. However, in our study, no additional synergistic effects were observed when CRTH2 antagonists were used as add-on treatment to corticosteroids. With consideration of the meta-regression analysis, which indicated no association between pre-bronchodilator FEV 1 and either asthma severity, concomitant treatment, or treatment duration, the non-superiority of add-on treatment of CRTH2 antagonists to corticosteroids might result from: 1) the difference in CRTH2 antagonists types and doses with various bioavailability, pharmacokinetics and pharmacodynamics; 2) the true benefit of CRTH2 antagonists being covered by the potent effects of concurrent corticosteroids use. For example, Hall et al. [14] also found that in steroids-naïve rather than steroids-on-use asthmatics 400 mg of BI 671800 could improve lung function [25].
In terms of PEF, our pooled and subgroup analysis revealed no improvement of morning and evening PEF in asthmatics with CRTH2 antagonists treatment. Moreover, one study [17] even reported that CRTH2 antagonists together with corticosteroids could reduce morning PEF. The poor relationship between FEV 1 and PEF might relate to the disassociation of the effect of CRTH2 antagonists on these two parameters [32,33]. However, the effect of CRTH2 antagonists on PEF should be  It is recommended by GINA that asthma assessment should focus on asthma symptom control and future exacerbations risk reduction [2]. ACQ and AQLQ are both commonly used self-evaluation questionnaires for asthma symptoms and quality of life [34,35], and our meta-analysis showed that CRTH2 antagonists therapy could reduce ACQ scores and increase AQLQ scores. SABA is one of the most important quick relievers for asthma onset, and the number of its rescue use has already been elucidated to be associated with asthma exacerbations [2]. Our meta-analysis also found that CRTH2 antagonists, either used as monotherapy or add-on therapy to corticosteroids, could reduce SABA use. Asthma exacerbations are associated with the poor asthma control [2] and is the major cause of morbidity and mortality in asthmatics [36]. In our meta-analysis, significant reduction in asthma exacerbations was found in the asthmatics with CRTH2 antagonist monotherapy rather than add-on therapy to corticosteroids. Therefore, based on the above positive findings, CRTH2 antagonists may serve as an efficacious surrogate for corticosteroids and reduce the use or adverse events of corticosteroids. However, future studies are still warranted due to the inconsistent of CRTH2 antagonists types, doses, and durations, as well as the potential heterogeneities and limited studies.
It is reported that sputum or blood eosinophil level is associated with high incidence of asthma attacks, and they are also one of the important markers for asthma phenotyping [37,38]. Meanwhile, blockade of CRTH2 has been recognized to down-regulate Th2 cytokines production [8], decrease eosinophils release from bone marrow [7,39], chemotaxis and respiratory burst [28]. Therefore, eosinophil might be a potential indicator for treatment effectiveness and asthma phenotyping may also help to identify the better responsive subgroup. However, inconsistent data from the included studies disabled us to pool in meta-analysis, and our systematic review also showed inconclusive results. FeNO is believed to be an indirect marker for eosinophilic airway inflammation [40], and our systematic review found that CRTH2 antagonists could not decrease FeNO. However, more studies are necessitated before we can draw a clear conclusion because the role of FeNO itself in asthma airway inflammation is still not clarified and controversial     findings showed that the specific inhibition of inducible nitric oxide synthase did not affect airway hyperresponsiveness and airway inflammation [41].
In our meta-analysis, we found similar adverse events between CRTH2 antagonists treatment and placebo, and no treatment related severe adverse events and deaths were reported, which indicated a general safety profile of CRTH2 antagonists in the treatment of asthma patients. Although CRTH2 antagonists were reported to cause some adverse events, but most of them were mild and moderate such as nasopharyngitis, headache, asthma, infections and gastrointestinal disorders. However, use with cautions, especially for some elderly patients with concomitant diseases, should always be addressed.
Several potential limitations require consideration in interpreting our study results. First of all, although some parameters, such as pre-bronchodilator FEV 1 , ACQ and AQLQ scores, have been improved in patients with CRTH2 antagonists, the clinical importance of these improvements need to be questioned because they are less than minimal clinical importance difference [35,42,43]. Secondly, asthma exacerbations in the trials included were not defined consistently and even not defined explicitly in one trial [22]. Thirdly, a small scale of some studies and limited number of RCTs included in several outcomes analysis may affect the power to explore the real outcome. Finally, the heterogeneities among the studies might cause inaccurate results in some outcomes. Although we have classified the studies into subgroups based on the intervention therapies and we found no statistical heterogeneities in most of the outcomes, but the baseline asthma severity and phenotypes varied among studies, which makes it necessary for further studies to clarify which subgroups of asthmatics can benefit this treatment. Moreover, given the variety of CRTH2 antagonists in selectivity, specificity and affinity, such as the dual affinity of AMG 853 to both DP2 and DP1, the interpretation of our results should also be cautious and it is hard to decide the optimal types of CRTH2 antagonists, dose, and treatment duration. Therefore, future studies involving and dealing with these issues are urgently needed.
Publisher's Note