Association between Dry Eye Disease, Air Pollution and Weather Changes in Taiwan

Abstract

Dry eye disease (DED) has become a common eye disease in recent years and appears to be influenced by environmental factors. This study aimed to examine the association between the first occurrence of DED, air pollution and weather changes in Taiwan. We used the systematic sampling cohort database containing 1,000,000 insureds of the National Health Insurance of Taiwan from 2004 to 2013, and identified a total of 25,818 eligible DED subjects. Environmental data, including those of air pollutants, temperature and relative humidity, were retrieved from the environmental monitoring stations adjacent to subjects' locations of clinics as exposure information. We applied the case-crossover design, which used the same subjects experiencing exposures on diagnosis days as cases and those on other days as controls. The descriptive statistics showed that the first occurrences of DED were the most for the elderly by age (53.6%), women by gender (68.9%), and spring by season (25.9%). Multivariate conditional logistic regression analyses indicated that carbon monoxide (CO), nitrogen dioxide (NO₂), and temperature were positively associated with DED (p < 0.05), while relative humidity was negatively related (p < 0.001). Because CO and NO₂ together are considered a surrogate of traffic emission, which is easier to control than the uprising temperature, it is suggested that efficient management and control of traffic emission may lower the probability of DED occurrence.

Keywords: Taiwan; air pollution; dry eye disease; relative humidity; temperature; traffic emission.

Figure 1

Annual mean concentrations of gaseous air pollutants (CO, NO₂, O₃, SO₂) from 2004 to 2013 (error bars denote standard deviations), derived from 73 of 76 monitoring sites.

Figure 2

Annual averages of particulate matter with aerodynamic diameter ≤2.5 and 10 µm (PM_2.5 and PM₁₀) concentrations, relative humidity (RH) and temperature from 2004 to 2013 (error bars denote standard deviations), derived from 73 of 76 monitoring sites.