BackgroundAllergic rhinitis is a common disorder that affects 10% to 40% of the population worldwide. Allergen immunotherapy (AIT) represents the only therapy that has the potential to resolve clinical symptoms of allergic rhinitis. However, up to 30% of patients do not respond to AIT. Biomarkers predicting the clinical efficacy of AIT as early as possible would significantly improve the patient selection and reduce unnecessary societal costs. MethodsArtemisia pollen allergic patients who received at least 1-year AIT were enrolled. Clinical responses before and after 1-year AIT were evaluated to determine AIT responders. Artemisia specific IgE and IgG4 levels were measured by using ImmunoCAP and enzyme-linked immunosorbent assay (ELISA) separately. Stepwise regression analysis was performed to identify which rhinitis-relevant parameters explained the most variability in AIT results. Liquid chromatography-tandem mass spectrometry (LC-MS/MS)-based proteomics was applied to identify the potential candidate biomarkers in the sera of responders and non-responders collected before and after 1-year therapy. The diagnostic performance of the potential biomarkers was then assessed using enzyme-linked immunosorbent assay (ELISA) in 30 responders and 15 non-responders.ResultsArtemisia specific IgE and IgG4 levels were elevated only in the responders. Regression analysis of allergic rhinitis-relevant parameters provided a robust model that included two most significant variables (sneeze and nasal congestion). Thirteen candidate biomarkers were identified for predicting AIT outcomes. Based on their association with allergy and protein fold change (more than 1.1 or less than 0.9), four proteins were identified to be potential biomarkers for predicting effective AIT. However, further ELISA revealed that only leukotriene A4 hydrolase (LTA4H) was consistent with the proteomics data. The LTA4H level in responders increased significantly (P < 0.001) after 1-year therapy, while that of non-responders remained unchanged. Assessment of LTA4H generated area under curve (AUC) value of 0.844 (95% confidence interval: 0.727 to 0.962; P < 0.05) in distinguishing responders from the non-responders, suggesting that serum LTA4H might be a potential biomarker for predicting the efficiency of AIT.ConclusionSerum LTA4H may be a potential biomarker for early prediction of an effective AIT.