Extracellular vesicles as biomarkers and disease mediators in atopic dermatitis: systematic review & meta-analysis

Background Atopic dermatitis (AD), more commonly known as eczema, is a chronic inflammatory disease that presents heterogeneously but is primarily characterized by dry skin, itching, erythema, and lichenification. Current diagnosis and severity assessment rely primarily on clinical evaluation, highlighting the need for objective biomarkers that more accurately reflect disease pathophysiology and treatment response. Extracellular vesicles (EVs) are membrane-bound particles that carry a variety of molecular cargo (such as proteins, DNA, RNA, and lipids) that often reflect the physiological state of their cellular origin. As a result, EVs are being increasingly studied as candidate biomarkers and therapeutics in a multitude of diseases. Methods A systematic search of PUBMED and Embase was conducted from database inception to August 9th, 2026. Studies involving human subjects or human-derived cell models that evaluated EVs as diagnostic, predictive, functional, or mechanistic biomarkers in AD were included. Study quality was assessed using a modified Newcastle-Ottawa Scale, and diagnostic performance was analyzed using receiver operating characteristic (ROC) curves and BRMA models when sufficient data were available. Results Eleven articles met the inclusion criteria. These studies included 284 individuals with AD (weighted mean age: 15.4 years 52.7% female) and a total of 187 controls (weighted mean age: 15.2 years; 48.7% female). Several candidate biomarkers demonstrated strong diagnostic performance, including EV-associated CD63, total filaggrin (FLG), 2x FLG, guaA, glmS, HMPREF0675_4995, and tRF-28-QSZ34KRQ590K. Functional studies suggested that circulating and bacteria-derived EVs contribute to keratinocyte dysfunction, inflammatory signaling, and epidermal barrier impairment, while a longitudinal study demonstrated treatment-associated changes in microbial EV-associated DNA profiles. Conclusion EV-associated RNAs, proteins, and microbial EV cargo represent biomarker candidates that may improve diagnosis, patient stratification, and understanding of AD pathophysiology. Future studies should validate these findings in larger longitudinal cohorts.