Abstract
Purpose :
To date, studies that linked dry eye disease (DED) to medication use were hypothesis-driven and did not investigate individual drugs. The purpose of this large study was to investigate the association between DED and both medication classes and individual drugs, using a hypothesis-free approach.
Methods :
79,866 participants (age range 20-97 yrs, 58% female) from the population-based Lifelines cohort in the Netherlands were cross-sectionally assessed for dry eye using the WHS dry eye questionnaire. Frequency of dryness and irritation symptoms were both scored; 0 (never), 1 (sometimes), 2 (often), or 3 (constantly) and a total of 2 or higher was assigned as symptomatic DED. All medications used were coded with the ATC classification system. Logistic regression analysis, adjusted for age and sex, was used to assess the risk of the 62 most used therapeutic/pharmacological subgroups (3rd level ATC) and the 100 most used individual drugs (5th level ATC) on symptomatic DED.
Results :
A total of 38 (61%) medication subgroups and 52 (52%) individual drugs were associated with dry eye (P<0.05). A multivariable model, including all medications that were univariably associated (P<0.10), but excluding medications that are used to treat DED or diseases known to cause DED (e.g. artificial tears and antirheumatic drugs), revealed the highest risk of DED (see Table) for antiglaucoma drugs, drugs for functional gastrointestinal disorders, psychostimulants for ADHD, drugs for peptic ulcer, drugs for constipation and urologicals. Low-ceiling diuretics appeared protective of dry eye symptoms. Using a similar approach, individual drugs that showed the highest risk of DED were mebeverine, methylphenidate, pantoprazole, omeprazole, isphagula, and risedronic acid. Hydrochlorthiazide and the oral contraceptive desogestrel and ethinylestradiol were associated with fewer dry eye symptoms.
Conclusions :
This large hypothesis-free study found medication use to be highly associated with DED and underlines the importance for doctors to assess medication status in DED. We discovered several new associations, confirmed but also debunked suggested associations from smaller studies, and found novel medication class differences. These results are highly relevant for potential treatment options and prevention strategies in DED.
This is an abstract that was submitted for the 2018 ARVO Annual Meeting, held in Honolulu, Hawaii, April 29 - May 3, 2018.