1. Introduction
As the global aging population continues to increase, the prevalence of age-related macular degeneration (AMD) has also risen as a consequence. In fact, AMD is the leading cause of visual impairment and loss of vision for the elderly [
1]. Wong et al. [
2] showed that 8.7% of the population has AMD, and the projected number of people worldwide with AMD will increase from 196 million in 2020 to 288 million in 2040. There are two types of AMD. One is dry (also called nonexudative) AMD, and the other is wet (also called exudative) AMD. Although the latter only accounts for 10–15% of all AMD cases, wet AMD can lead to severe visual impairment and blindness, dramatically affecting an individual’s independence and quality of life [
1,
3,
4,
5,
6,
7]. Moreover, AMD treatments are costly and may not be affordable to everyone in both developed and developing countries, thus potentially imposing a substantial burden on the healthcare sector, society, and global economy. The risk factors of AMD include age, race, smoking, and diet [
1], while hypertension and diabetes are found to potentially relate to AMD development [
1,
6].
Another important factor that affects AMD is ambient air pollution. Exposure to air pollution is associated with AMD [
8]. Ambient air pollution, such as fine particulate matter (PM
2.5), increases the risk of AMD [
9], while long-term exposures to PM
10, NO
2, and CO may raise its prevalence [
10]. Since air pollution can induce oxidative stress and inflammation, it is likely to escalate the incidence of AMD [
8,
9], because the retina is one of the most oxygen-consuming tissues in the human body due to the need for metabolism. When the accumulation of reactive oxygen species is unbalanced, it can lead to oxidative stress; thus, air pollution causes oxidative stress and consequent inflammation [
9,
10].
Ambient air pollution can come from cars, factories, forest fires, and dust storms (DSs). DSs often arise in the Middle East, Central Asia, North America, Central Africa, and Australia. In Taiwan, DSs originating from the deserts of Mongolia and China, can affect the western part of the island during the winter and spring [
11]. The pollutants in the dust combined with intensified desertification in China are resulting in more and longer DS events [
12,
13]. Particulate matter (PM) concentration levels also increase due to DS [
13,
14,
15,
16,
17].
Past studies have identified associations between DSs and hospitalization due to respiratory and cardiovascular diseases [
11,
12,
13,
14,
15,
16,
18,
19,
20]. However, the effects of DSs on the eye, which is directly exposed, are rarely discussed. Only two studies showed that DSs impact a person’s eyes. Chien et al. [
21] reveal significantly acute impacts on children’s conjunctivitis clinic visits during DS periods, especially school children. Mu et al. [
22] indicated that the occurrence of eye lacrimation is related to DSs. To the best of our knowledge, there is no study analyzing the relationships between DSs and AMD. Using a 5 year population-based and representative national dataset, the main purpose of this study was to fill the gap in the literature by examining the association of DS events with the daily number of AMD hospital admissions and outpatient visits in Taiwan.
2. Materials and Methods
2.1. Data
AMD outpatient visits and hospitalizations during 2008–2012 were retrieved from a representative sample of longitudinal health insurance data with one million cases built upon the population-based National Health Insurance Research Database (NHIRD) in Taiwan. AMD can be divided into dry AMD and wet AMD. Dry AMD is a chronic atrophic eye disease that can lead to blindness after decades. This type accounts for about 85–90% of overall cases. The diagnosis codes (ICD-9-CM) are 362.51 and 362.57. Wet AMD is a neovascular disease that may cause vision loss over a period of weeks to months and requires urgent treatment. This type accounts for about 10–15% of overall cases. The diagnostic codes (ICD-9-CM) are 362.52 and 362.53 [
23,
24].
Air pollution data were taken from Taiwan’s Environmental Protection Administration (TEPA) database. TEPA defines DS events using the PM10 concentrations measured by five indicator observation stations, including Yilan, Wanli, Guanyin, Matsu, and Yang Ming. DSs in Taiwan mainly come from overseas areas, and the locations of these indicator observation stations are all on offshore islands or in northern coastal areas. Therefore, the measured concentrations are not easily affected by local stationary sources of air pollution and mobile sources of air pollution. Normally, the 24 h moving average concentration of PM10 is about 50 μg·m−3, but PM10 can rise to above 100 μg·m−3 when DSs emanate from overseas. TEPA uses 100 μg·m−3 as the criterion (a total of 11 DS events occurred between 2008 and 2012) for measurement.
The main substances of air pollution include PM10, PM2.5, CO, O3, NO2, and SO2. When DSs occur, both PM10 and PM2.5 increase simultaneously, because the concentrations of these particles are included in the DS data. Therefore, this study used other air pollutants to represent the impact of air pollution. According to the monitoring data of TEPA, only the concentrations of PM10, PM2.5, and O3 reached unhealthy levels in recent years, while the other substances were at good levels; hence, this paper used O3 as an explanatory variable.
According to previous studies, the concentration of NO
2 has an important impact on AMD; thus, it was also included in the research model [
8,
9,
10]. The concentrations of O
3 and NO
2 were calculated from daily averages of 60 air quality observation stations across Taiwan. Weather data were taken from Taiwan’s Central Weather Bureau database. The ambient temperature was calculated from the daily average temperature of 23 meteorological observation stations across Taiwan.
2.2. Methods
This study employed the Poisson time-series analysis proposed by Brännäs [
25,
26]. The main reason is that the data of daily outpatient visits and hospitalizations are aggregated time series of count data. We separated the samples by disease type and severity. AMD was divided into dry AMD and wet AMD, as well as inpatient AMD and outpatient AMD. In addition, according to past studies, gender and age are also important factors affecting AMD [
27,
28]. Therefore, samples were stratified according to gender (male and female) and age groups (below 50, 51–60, 61–70, 71–80, and over 81). The empirical model is defined below.
The dependent variable, AMDt, is the daily number of AMD outpatient visits or hospitalizations; α0 is the intercept term; t is the time trend; Dayi is a dummy variable for DS events. The impact of DSs on AMD mainly occurs through particulate matter (PM). Particulate matter enters the lungs from the respiratory tract and then enters the blood circulation from the pulmonary blood vessels. Particulate matter flowing through the eyes can lead to AMD. This process can take days to have an effect; therefore, we used Day0 to capture the immediate effects and Day1 to Day3 to capture the delayed effects (with no-DS days as the reference group). AMDt−j is the autoregressive term of AMD. We selected the optimal autoregressive lags on the basis of the Ljung–Box test and autocorrelation function (ACF) in order to minimize the autocorrelation in error. Xt represents other confounders, including atmospheric data such as ambient temperature, NO2, O3, and season (with summer as the reference group). Lastly, εt is the error term.
4. Discussion
Air pollution is associated with respiratory and cardiovascular diseases such as stroke, asthma, and pneumonia [
11,
12,
13,
14,
15,
16,
29]. Since eyes are directly exposed to air pollutants, it is likely that air pollution can cause eye-related diseases. However, only a few studies focused on examining the impact of air pollution on eye diseases [
10,
30]. One of the major sources of air pollution comes from DSs, but research on the correlation between DSs and eye diseases is rare. There are only two studies that explored the impact of DSs on conjunctivitis and eye lacrimation [
21,
22]. AMD is one of the leading eye diseases that causes blindness in the elderly [
1]. However, no examination has been conducted for the association between DSs and AMD. Therefore, this study examined the associations between DS events and AMD healthcare using a large national health insurance research database from 2008 to 2012 in Taiwan.
The present study observed that AMD outpatient visits and hospitalizations were associated with DS events, yet the effect was not immediate, but delayed. Our results indicated that DS event days did not result in a significant higher number of AMD outpatient and inpatient cases. However, a significantly higher number of AMD cases were observed 1 and 2 days after DS events. Since DSs can increase PM concentration levels, the delayed effect may be explained by the delayed biological effects following exposure to residual PM after DS events [
12,
29]. An increase in PM concentration levels initiates systemic oxidative stress and consequent lipid peroxidation, which activates the innate immune system and increases inflammation in the retina and cells, potentially resulting in increased risk of AMD [
10]. This process takes some days. Similar delayed effects can also be found in the literature regarding the impact of DS events on other diseases [
11,
12,
13,
19]. Consistent with previous studies that examined the relationships between ambient air pollution and AMD [
8,
9,
10], this study confirmed a positive relationship between DSs and AMD. Our results are also consistent with Chien et al. [
21] and Mu et al. [
22], who showed that the prevalence of eye diseases increased after DS events.
On the basis of the total population, we compared the difference between wet AMD and dry AMD results and found that only dry (but not wet) AMD outpatient and inpatient numbers were significantly higher 1 and 2 days after DSs. However, when the data were stratified by gender, we found that wet outpatient visits and hospitalizations 1 and 2 days after DSs had a significantly higher number than no-event days for females. A significantly higher number of dry AMD cases and hospitalizations was found 1 day after DSs for males only. In other words, women were more likely to undergo more serious wet AMD outpatient and inpatient care than men following DS events. Some previous studies indicated a relatively higher prevalence of wet AMD in women. Rudnicka et al. [
31] showed a higher risk (OR = 1.2) of wet AMD in women compared with men. A study conducted in southwestern Taiwan by Huang et al. [
32] also noted that the prevalence of early AMD was higher in men than in women.
When stratifying the data into different age groups, the results indicated that the impact of DS events on AMD across all age groups was comprehensive. AMD is the leading cause of irreversible blindness in people above age 50 years old [
33,
34], and age is by far the most significant factor for AMD [
1]. The prevalence of AMD is gradually increasing as a consequence of exponential population aging [
2]. It is not surprising to find that numbers of both wet AMD outpatient and dry AMD inpatient visits were significantly higher 1 day after DSs than on no-event days for the groups over 81 years old. People aged 71–80 and 51–60 also tended to have a significantly more dry AMD outpatient visits 2 days after DSs and wet AMD hospitalizations 1 day after DSs, respectively. Surprisingly, a positive and significant association between dry AMD outpatient visits 2 days after DSs appeared among people under 50 years old. Even though dry AMD is the most common type of AMD, incurring almost no vision loss, it can progress to wet AMD [
4,
5,
6]. However, a younger age of getting dry AMD results in a sooner transition into wet AMD and a subsequent risk of visual impairment. People under 50 still have to be careful about DSs and AMD, despite the literature suggesting that AMD is more common in people 50 years of age or older.
Our results also showed a significantly positive association between temperature and pollutants and AMD healthcare. High temperatures tend to cause more dry and wet AMD care, especially for males and the elderly. High amounts of NO2 and O3 are also likely to cause increased dry and wet AMD cases. Lastly, the overall trend in AMD outpatient care increased, except for the <50 and 71–80 age groups in dry AMD outpatient care. A growing trend in wet AMD hospitalizations was also found in the ≥81 age group. Conversely, dry AMD hospitalizations trended downward regardless of gender and age group.
The strengths of this research include the use of a large sample size and highly accurate AMD medical data from the NHI research database with information on DSs, meteorology, and air pollution. Our findings provide evidence that the association between AMD and DS does exist and add to the growing evidence of the damaging effects of ambient air pollution on AMD. However, some limitations existed in our study. First, personal exposure levels to DS events were unknown. We used NHI data to identify patients with AMD healthcare use without knowing if this was because of DS events. Second, our AMD data were aggregated daily medical data. Personal risk factors for AMD such as smoking, BMI, and hypertension, therefore, could not be controlled for in the study. Third, despite the increase in daily AMD cases following DS, it should be remembered that the total number of AMD cases associated with DS were still much lower than the overall AMD cases, as DSs only occurred 11 times over a 5 year period.