PURPOSE: To test the association between serum inflammatory markers and dry eye disease (DED) using a hypothesis-free proteomic approach in a population-based cohort. METHODS: A total of 2602 unselected community-based participants (mean age 61.5 (range 21-92 years), 94.4 % female) from the TwinsUK cohort were examined. DED was assessed with the validated Women's Health Study (WHS) questionnaire
cases were defined by either a previous clinician diagnosis or presence of highly symptomatic dry eye. Serum inflammatory markers were assessed with the Olink Target 96 Inflammation panel. We performed logistic regression mixed effect models, adjusted for age, BMI, sex, and twin relatedness, with false discovery rate (FDR) correction. RESULTS: Prevalence of WHS-defined DED was 29.1 %, with 26.2 % having a previous diagnosis of DED and 16.5 % having highly symptomatic dry eye. Of 74 inflammatory markers, significant associations with WHS-defined DED were found for neurotrophin-3 (NT-3
OR: 0.68, FDR p-value: 0.043), natural killer-cell receptor 2B4 (CD244
OR: 0.68, FDR p-value: 0.043), C-X-C motif chemokines (CXCL) 9 (OR: 1.23, FDR p-value: 0.043) and CXCL10 (OR: 1.22, FDR p-value: 0.043). Significant association with highly symptomatic dry eye were found with increased levels of CCL19, CXCL9, CXCL10, CCL20, CX3CL1 (fractalkine), TNF, CDCP1, and CCL25. CONCLUSIONS: This large population-based study found several serum inflammatory proteins to be associated with DED, confirming and adding to previous targeted tear and corneal and conjunctival expression studies in murine models and clinic-based case-control studies. Of interest, a novel potential biomarker NT-3, which plays a role in corneal nerve function, was identified.