BACKGROUND: Chronic obstructive pulmonary disease (COPD) is now a major contributor to illness and death on a global scale. Traditional classifications based on pulmonary-function variables like the FEV1 (% predicted), lack sensitivity to early inflammatory changes and disease heterogeneity. In contrast, MiRNAs have emerged as promising biomarkers for respiratory conditions including COPD. The aim of this study was to examine the association between selected MiRNAs and COPD. METHODS: A total of 200 COPD patients (GOLD I-IV, n=50 each) and 50 healthy controls were enrolled. Differentially expressed miRNAs were identified from the GEO database and validated by RT-qPCR. Univariate analysis and LASSO regression were used for feature selection. A logistic regression model incorporating selected variables was established for the forecast COPD severity. KEGG pathway enrichment of miR-518b target genes was performed using DAVID. RESULTS: Five miRNAs (miR-216a, miR-518b, miR-106a, miR-1233, and miR-184) were substantially increased in COPD and correlated with disease severity (P<0.001). LASSO regression identified FEV1 (% predicted), DLCO (% predicted), CAT score, miR-518b, and age as key predictors. The combined model showed excellent classification performance (AUC=0.953; sensitivity=92.9%; specificity=81.3%). MiR-518b emerged as a strong independent risk factor (OR=18.91, P=0.003). Gene set enrichment of miR-518b targets pointed to involvement in both the Toll-like receptor and Hippo pathways, implicating its critical roles in inflammation and airway remodeling. CONCLUSIONS: MiR-518b is closely associated with COPD severity and may be useful in clinical practice. A model integrating miRNA expression and clinical parameters provides high predictive value for COPD classification and supports precision diagnosis.
Zhang et al. (Mon,) studied this question.