Association between Dry Eye Parameters Depends on Tear Components

Abstract

How tear components contribute to dry-eye symptoms/signs remains less well-defined. This observational cross-sectional study enrolled 4817 (F/M = 3590/1227) patients. Subjective symptoms were evaluated with the SPEED and OSDI questionnaires. Fluorescein tear breakup time (FTBUT), superficial punctate keratitis (SPK) grading, Schirmer scores, number of expressible meibomian glands (MGE), lipid layer thickness (LLT), blink/partial blink rates and meibography were recorded. Patients were divided into 4 types according to their Schirmer scores and LLT, i.e., Type 1 (N = 1494): Schirmer > 5 mm, LLT > 60 nm; Type 2 (N = 698): Schirmer > 5 mm, LLT ≤ 60 nm; Type 3 (N = 1160): Schirmer ≤ 5 mm, LLT ≤ 60 nm; Type 4 (N = 1465): Schirmer ≤ 5 mm, LLT > 60 nm. Lipid deficiency (LLT ≤ 60 nm) and aqueous deficiency (Schirmer score ≤ 5 mm) were found in 38.6% and 54.5% of patients, respectively. The majority (62.4%) of lipid-deficient patients were also aqueous deficient, while 44.2% of aqueous-deficient patients were also lipid-deficient. Type 3 patients (mixed type) had the highest symptom scores (p = 0.008 and 0.007 for SPEED and OSDI, respectively), more total blinks (p < 0.001) and the shortest FTBUT (p < 0.001). Stepwise multiple regression demonstrated that LLT and Schirmer score were significant contributors to FTBUT in all 4 types. The FTBUT correlated with SPK severity in all 4 types, with Schirmer score in types 1 and 4, and with LLT in type 3 patients. SPK correlated with LLT and MGE in types 1 and 4. Age correlated with dry eye parameters more significantly than sex. Subtyping by aqueous and lipid components facilitates the understanding of dry eye pathophysiology.