Correlation between diagnostic tests guides the development and use of a dry eye diagnostic algorithm

Abstract

Abstract Purpose: The existence of large variations in results of diagnostic tests for cases of mild to moderate dry eye is widely recognised. The purpose of this study was to assess if there was correlation between existing dry eye diagnostic tests and attempt to develop a diagnostic algorithm. Methods: A total of 91 subjects were recruited to the study. The tear film and ocular surface were evaluated by measuring tear volume using the phenol red thread test, tear break up time using fluorescein, biomicroscopic examination of meibomian glands and impression cytological (IC) assessment of conjunctival goblet cells. Dry eye symptoms were assessed using McMonnies dry eye questionnaire and statistical correlations were assessed between all dry eye tests used. Results: In this study cohort there were no severe aqueous deficient dry eye patients. Meibomian gland pathology, questionnaire (score of ≥14), IC and an altered tear break up time (reference ≤7 seconds) demonstrated correlation, while no other tests demonstrated evidence of correlation. A diagnostic algorithm was devised using these tests alone, with maximum weighting placed on questionnaire with decreasing weighting assigned to IC goblet cell grade, meibomian gland pathology and TBUT. Conclusions: We propose this development of a flexible probabilistic algorithm as a rational approach when diagnosing mild/moderate dry eye in large study groups. Until further studies have demonstrated the true sensitivities and specificities of dry eye tests within the context of test sequences and particular patient cohorts are known, comparison of results between dry eye studies will remain questionable.