Chronic rhinosinusitis (CRS) is a common comorbid condition of asthma that affects the long-term outcome of asthmatic patients. CRS is a heterogeneous disease requiring multiple biomarkers to explain its pathogenesis. This study aimed to develop potential biomarkers for predicting CRS in adult asthmatic patients in a real-world clinical setting.

This study enrolled 108 adult asthmatic patients who had maintained anti-asthmatic medications, including medium-to-high doses of inhaled corticosteroid plus long-acting β2-agonists, and compared clinical characteristics between patients with CRS (CRS group) and those without CRS (non-CRS group). CRS was diagnosed based on the results of paranasal sinus X-ray and/or osteomeatal-unit CT as well as clinical symptoms. Type-2 parameters, including blood eosinophil count, serum levels of periostin/dipeptidyl peptidase 10 (DPP10) and clinical parameters, such as FEV1% and fractional exhaled nitric oxide (FeNO), were analyzed. All biomarkers were evaluated by logistic regression and classification/regression tree (CRT) analyses.

The CRS group had higher blood eosinophil counts/FeNO levels and prevalence of aspirin-exacerbated respiratory disease (AERD) than the non-CRS group (n = 57, 52.8% vs. n = 75, 47.2%; P < 0.05), but no differences in sex/smoking status or asthma control status were noted. The CRS group had higher serum periostin/DPP10 levels than the non-CRS group. Moreover, logistic regression demonstrated that serum periostin/DPP10 and the AERD phenotype were significant factors for predicting CRS in asthmatic patients (adjusted odds ratio, 2.14/1.94/12.39). A diagnostic algorithm and the optimal cutoff values determined by CRT analysis were able to predict CRS with 86.27% sensitivity (a 0.17 negative likelihood ratio).

Serum periostin, DPP10 and the phenotype of AERD are valuable biomarkers for predicting CRS in adult asthmatic patients in clinical practice.