The repeatability of clinical measurements of dry eye.

PURPOSE: The repeatability of individual dry eye diagnostic tests has been reported in the literature on normal samples of patients and to a lesser degree in dry eye patients. In this study, the repeatability of a battery of clinical diagnostic tests for dry eye was assessed on mild to moderate dry eye patients.
METHODS: A dry eye examination was performed on 75 patients on two occasions by a single examiner. The battery of dry eye tests included symptom assessment, contact lens and medical history, slit-lamp biomicroscopic evaluation of the eyelids, evaluation of Meibomian glands, assessment of tear film quality, tear meniscus height, assessment of blink quality, fluorescein tear breakup time (TBUT), fluorescein and rose bengal staining of the cornea and conjunctiva, phenol red thread test, and Schirmer test.
RESULTS: The repeatability of subjective report of dryness (kappaW = 0.62) and grittiness (kappaW = 0.73) was moderate to high. In contrast, the repeatability of Meibomian gland disease classification (kappaW = 0.20), presence or absence of inferior corneal fluorescein staining (kappa = 0.25), and inferior conjunctival rose bengal staining (kappa = 0.21) was poor. When a summed staining score of corneal and conjunctival regions was evaluated, weighted kappa reliability was fair. The repeatability of tear breakup time was substantial [95% limits of agreement -5.71 to 5.83 seconds; intraclass coefficient coefficient (ICC) 0.65], and improved when the two timed readings were averaged. Repeatability of the Schirmer test is more variable as wetting scores increase. When the average Schirmer scores </= 10 mm were evaluated, moderate repeatability was demonstrated, indicating that the test performs better with more advanced disease.
CONCLUSION: Although patient-reported symptoms are moderately repeatable from visit to visit, many of the procedures clinically used to diagnose and monitor dry eye syndromes are largely unrepeatable.