In the present study, we consecutively enrolled 160 HDM-induced AR patients (AR group) who visited the Department of Otolaryngology Head and Neck Surgery, Xiangya Hospital of Central South University between the period of January 2017 and July 2017. Seventy-seven sex- and age-matched control subjects without any allergic diseases were recruited as the healthy control (HC) group. This study was conducted under the guidance of the Helsinki Declaration and approved by the Human Ethical Committee of Xiangya Hospital of Central South University, and all participants submitted written informed consent. HDM-induced AR was diagnosed by clinicians based on the allergic rhinitis and its impact on asthma (ARIA) guidelines (Brożek et al., 2017). The inclusion criteria were listed as follows: 1) persistent AR with typical allergic symptoms for more than 2 years, 2) a positive test result in nasal provocation, 3) a positive skin test to Dermatophagoides farinae (Der f) and/or Dermatophagoides pteronyssinus (Der p) (at least ++) and/or positive specific IgE against Der f and/or Der p (>0.35 IU/ml), and 4) age ≥18 years. The exclusion criteria were listed as follows: 1) active asthma that required inhaled corticosteroid treatment; 2) significant vasomotor rhinitis; 3) with other allergic diseases, inflammatory or septic diseases, or autoimmune diseases; 4) pregnant or planning pregnancy within the next three years; 5) severe cardiovascular diseases and impaired liver and renal function; 6) with an acute upper respiratory tract infection within 4 weeks; and 7) use of anti-allergic drugs within 4 weeks.

Patients were asked to record their disease symptoms with the visual analogue scale (VAS) and total nasal symptom score (TNSS) as previously described (Adamko et al., 2018; Zhao et al., 2019b). VAS is a 10-point scale from 0 (no symptom) to 10 (maximum). TNSS is the sum of four individual symptom scores (sneezing, rhinorrhea, nasal congestion, and nasal itching), and each symptom ranged from 0 to 3, resulting in a possible score of 0–12 (detailed described in Supplementary Table S1). Based on TNSS, AR patients were categorized into the mild AR (MAR) group (TNSS ≤4) and the moderate–severe AR (MSAR) group (TNSS >4) according to recommendation by previous publications (Refaat et al., 2015; Adamko et al., 2018).

Five ml fresh venous blood were collected from AR patients and HCs before breakfast, and stored at room temperature for 1 h after sampling. Then, collected blood samples were centrifuged at 4°C (3,000 rpm for 10 min), and the supernatants were collected and stored at -80°C in 1.5 ml ep tubes (NEST Biotechnology, Wuxi, China) for subsequent experiments. All serum samples were coded before the measurements. Serum MIF concentrations were quantified with a commercial ELISA kit (Multisciences, Hangzhou, China) according to the manufacturer’s instructions. All serum samples were diluted 1:40 and run in duplicate to improve assay precision.

In the MSAR group, 106 patients voluntarily received HDM SLIT for a period of 3 years to obtain long-term efficacy as previously suggested (Li et al., 2019). Eligible participants were administrated with the standardized Der f drops which were provided by Wolwopharma Biotechnology Company (Zhejiang, China) according to the schedules recommended by the manufacturer. SLIT consists of a step-up dosage phase and a maintenance phase, the detailed schedules were conducted as described in our previous study (Xie et al., 2021), and all adverse reactions were recorded.

As patient compliance was an important factor affecting the efficacy of SLIT, all participants received patient education to improve their adherence as previously recommended, including initial education with publicity brochures (explaining the course, costs, efficacy, and safety of SLIT) at the first two months during the SLIT, and follow-up education through department visits or telephone (assessing patient’s recent clinical symptoms, adjusting the SLIT schedules, and guiding to avoid and manage adverse reactions) (Kiotseridis et al., 2018; Li et al., 2019). As SLIT cannot completely alleviate and even aggravate allergic symptoms during the treatment especially when allergen load is heavy, appropriate pharmacotherapy was provided to patients, including antihistamine and corticosteroid (Canonica et al., 2007; Li et al., 2019), and the medication scores (MSs) were recorded. The definition of MSs is the sum of medication consumption in the last week and scores according to the World Allergy Organization recommendations: oral or intranasal antihistamines = 1 point; nasal glucocorticoids = 2 point; and oral glucocorticoids = 3 point (Pfaar et al., 2014). The symptom and medication score (SMS) was calculated as the sum of TNSS and MS/7, and the efficacy of SLIT was evaluated on the basic of SMS as described in our previous publication; a change of >30% point compared to the baseline SMS level represents good response; otherwise, the SLIT was defined as poor response (Xie et al., 2020; Xie et al., 2021).

Normally distributed data were shown as mean ± standard deviation (SD), and non-normal distributed data were described as median and interquartile range. One-way analysis of variance (ANOVA) or the Mann–Whitney U test was utilized for comparison among three groups, and the subsequent comparisons were performed by Student–Newman–Keuls (SNK) to locate the source of significances; Student's t test or the Kruskal–Wallis H test was used for comparison between two groups. Categorical data were expressed as number (%) and compared utilizing the chi-square test. To investigate the correlations among MIF level, TNSS, VAS, and other clinical variables, Spearman’s test was conducted. The multivariate logistic regression analysis was performed to identify the independent predicting factors for the clinical responsiveness of SLIT. A receiver operating characteristic (ROC) curve was constructed to determine the utility of serum MIF and other indications as markers for predicting the efficacy of SLIT in given specimens, and the area under the curve (AUC), sensitivity, specificity, and cutoff value were evaluated. For all tests, p < 0.05 was considered statistically significant. All statistical analyses and ROC analysis were conducted by SPSS Statistics for Windows (Version 25.0, IBM Corp., Armonk, NY), and other figures were constructed in GraphPad Prism (Version 7.0, Software Inc. La Jolla, CA, United States).

The basic demographic and clinical characteristics of all participants are displayed in Table 1. Age, sex, disease duration, BMI, and smoking were similar among three groups (p > 0.05). Patients in the MSAR group showed higher levels of serum total IgE, HDM-specific IgE, blood eosinophil count and percentage, TNSS, and VAS (all p < 0.05) than those in the HC and MAR groups.

As presented in Figure 1, the serum MIF levels were 20.7 ± 9.2 ng/ml in the AR group, which were markedly higher than those in the HC group (12.8 ± 4.7 ng/ml, p < 0.001), and MIF levels were significantly greater in the MSAR group (21.9 ± 9.8 ng/ml) than those in the MAR group (17.3 ± 6.5 ng/ml) and the HC group (12.8 ± 4.7 ng/ml, all p < 0.05). Spearman’s correlation analysis results showed that the serum levels of MIF in AR patients are positively correlated with TNSS (r = 0.599, p < 0.001), VAS (r = 0.505, p < 0.001), serum HDM–specific IgE (r = 0.7519, p < 0.001), total IgE (r = 0.221, p = 0.005), blood eosinophil count (r = 0.5639, p < 0.001), and percentage (r = 0.7589, p < 0.001; Figure 2).

In the present study, 106 MSAR patients received SLIT, and 26 patients dropped out because of various reasons (Table 2), and the serum MIF levels in these 29 patients were 21.4 ± 7.9 ng/ml. During the treatment, 13 patients reported minor adverse events (Table 3) and no patient complained of major systemic adverse reactions. Finally, 80 MSAR patients successfully completed the whole course of SLIT and provided valid follow-up data: 45 patients were categorized into the good-response group and other 35 cases were grouped into the poor-response group. As Table 4 shows, the serum HDM–specific IgE, blood eosinophil count, and percentage in the good response group were significantly lower than those in the poor-response group (all p < 0.05), but no statistical difference was found in age, sex, disease duration, BMI, smoking, serum total IgE, TNSS, and VAS. The concentrations of serum MIF were markedly lower in the good-response group (17.9 ± 7.1 ng/ml) than those in the poor-response group (23.5 ± 8.8 ng/ml, p < 0.001; Figure 3). The ROC analysis results in Figure 4 showed serum MIF exhibited better accuracy and utility to predict clinical response of SLIT in AR patients than serum HDM–specific IgE, blood eosinophil count, and percentage. The detailed parameters of ROC analysis are displayed in Table 5. Based on the optimal cutoff values of the serum MIF level obtained by ROC analysis, patients were categorized into low MIF level group (serum MIF level <22.5 ng/ml, n = 51) and high MIF level group (serum MIF level >22.5 ng/ml, n = 29) as similarly described in previous publications (Liang et al., 2020; Wu et al., 2021). In low MIF level group, the good-response rate was 78.4%, which was higher than that in high MIF level group (17.2%, p < 0.001). In addition, the serum HDM–specific IgE, blood eosinophil count, and percentage were significantly lower in low MIF level group than those in the high MIF level group (all p < 0.05; Figure 5)

Table 4 revealed the results of the possible factors associated with clinical response of SLIT in the univariate analysis, and variables with a p value <0.05 in the univariate analysis and serum MIF were further included in the multivariate logistic regression analysis. Unadjusted and adjusted multivariate analysis models both demonstrated that serum MIF was an independent factor associated with the clinical efficacy of SLIT (p < 0.001) (Table 6).