Article Text

Download PDFPDF

Traffic-related air pollutants increase the risk for age-related macular degeneration
  1. Kuang-Hsi Chang1,2,3,
  2. Po-Yuan Hsu4,
  3. Chun-Ju Lin5,6,
  4. Cheng-Li Lin7,
  5. Suh-Hang Hank Juo3,4,8,9,10,
  6. Chung-Ling Liang10,11,12,13
  1. 1 Department of Medical Research, Tungs' Taichung Metroharbor Hospital, Taichung, Taiwan
  2. 2 General Education Center, Jen-Teh Junior College of Medicine, Nursing and Management, Miaoli, Taiwan
  3. 3 Graduate Institute of Biomedical Sciences, China Medical University, Taichung, Taiwan
  4. 4 Department of Medical Research, China Medical University Hospital, Taichung, Taiwan
  5. 5 Department of Ophthalmology, China Medical University Hospital, Taichung City, Taiwan
  6. 6 School of Medicine, College of Medicine, China Medical University, Taichung, Taiwan
  7. 7 Management Office for Health Data, China Medical University Hospital, Taichung, Taiwan
  8. 8 Institute of New Drug Development, China Medical University, Taichung, Taiwan
  9. 9 Drug Development Center, China Medical University, Taichung, Taiwan
  10. 10 Center for Myopia and Eye Disease, China Medical University Hospital, Taichung, Taiwan
  11. 11 Department of Ophthalmology, Asia University Hospital, Taichung, Taiwan
  12. 12 Department of Optometry, College of Medical and Health Science, Asia University, Taichung, Taiwan
  13. 13 Bright Eyes Clinic, Kaohsiung, Taiwan
  1. Correspondence to Professor Suh-Hang Hank Juo; hjuo{at} and Dr. Chung-Ling Liang; ling0228{at}


The aim of this study was to investigate whether ambient nitrogen dioxide (NO2) and carbon monoxide (CO) increase the risk for age-related macular degeneration (AMD). This is a longitudinal population-based study using the data on Taiwan National Health Insurance Program between year 2000 and 2010. From the nationwide dataset, we enrolled subjects aged 50 or older and the annually total NO2 and CO exposure was calculated from 1998 to 2010 for each subject. The Cox proportional hazard regression was used to estimate the HRs with adjustment for other variables. A total of 39,819 AMD-free residents were enrolled, and 1442 participants developed AMD during the 11 -year follow-up. Compared with the lowest exposure quartile, the highest quartile of each air pollutant was associated with an increased risk for AMD. The adjusted HR was 1.91 (95% CI 1.64 to 2.23, p<0.001) for the highest NO2 quartile, and was 1.84 (95% CI 1.5 to 2.15, p<0.001) for the highest CO quartile. In this study, chronic exposure to the highest quartile of ambient NO2 or CO significantly increases the risk for AMD.

  • traffic-related pollutants
  • nitrogen dioxide (NO2)
  • carbon monoxide (CO)
  • age-related macular degeneration (AMD)

Statistics from

Significance of this study

What is already known about this subject?

  • Chronic exposure to air pollution is associated with a high risk for several diseases including respiratory and cardiovascular diseases.

  • Previous studies have shown that air pollution increases a risk for conjunctivitis.

  • Age-related macular degeneration (AMD) is a late-onset disease characterized by the formation of lipid-rich extracellular deposits, localized inflammation, and ultimately neurodegeneration in the central part of the retina.

  • AMD can be caused by interactions between genetic and environmental risk factors.

What are the new findings?

  • The adjusted HR was 1.91 (95% CI 1.64 to 2.23, p<0.001) in the highest nitrogen dioxide (NO2) exposure group compared with lowest NO2 exposure group.

  • The adjusted HR was 1.84 (95% CI 1.57 to 2.15, p<0.001) in the highest carbon monoxide (CO) exposure group compared with lowest CO exposure group.

  • The trend tests for increased risk of AMD were significant in both NO2 and CO exposure.

How might these results change the focus of research or clinical practice?

  • This study indicates the air pollution exposure as a risk factor for AMD.

  • Further animal studies are necessary to evaluate the association between air pollutants and macular degeneration.


With the rapid industrialization and urbanization in many countries, the air quality problem caused by heavy traffic has become an important issue for public health. Chronic exposure to air pollution is associated with a high risk for several diseases including respiratory and cardiovascular diseases.1–4 Although the effect of air pollution on eye disorders has been less studied, previous studies have shown that air pollution increases a risk for conjunctivitis and dry eye.5 6 Because the cornea and conjunctiva are directly exposed to the outdoor air, it is reasonable to conceive that the outer segment of eyes is more susceptible to air pollution. However, recent studies have indicated that air pollution also induces systemic diseases including dementia.7–9 Therefore, the inner segment of the eye may be also affected by air pollution.

Age-related macular degeneration (AMD) is a late-onset disease characterized by the formation of lipid-rich extracellular deposits, localized inflammation, and ultimately neurodegeneration in the central part of the retina (termed the macula). AMD is one of the major causes of irreversible visual impairment among the elderly in the developed world.10 11 AMD has increasingly become a global issue and the Asian population over the age of 50 years has high incidence of AMD.12 As the aging population increases, the prevalence of AMD is expected to increase proportionately. Several risk factors have been proposed for AMD and it is likely that AMD can be caused by interactions between genetic and environmental risk factors.13 Studies also reveal a link between cardiovascular disease and AMD.14 15

Given that air pollution can increase a risk for cardiovascular diseases and neurological diseases, it is biologically plausible to hypothesize that air pollution may also influence AMD development. We specifically focused on two major traffic pollutants (nitrogen dioxide (NO2) and carbon monoxide (CO)). To test our hypothesis, we analyzed data on a longitudinal cohort study using the Taiwan National Health Insurance data to assess the contribution of traffic-related air pollution to AMD.

Methods and materials

The present study used a nationwide longitudinal cohort to investigate the risk of AMD after chronic exposure to ambient NO2 and CO.

Data source

The present study used two databases, the Longitudinal Health Insurance Database (LHID) and Taiwan Air Quality Monitoring Database (TAQMD). LHID contained 1 million insurants randomly selected from the original Registry for beneficiaries joining in the Taiwan National Health Insurance Program. This insurance program was set up by the Taiwan Bureau of National Health Insurance since March 1995. At the end of 2009, this insurance covered over 99% Taiwan residents. LHID included all medical records from the start of 1996 to the end of 2010. The identification of insurant was re-coded before the data were released to researchers because of the Personal Information Protection Act. To identify the disease in LHID was according to the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM).

TAQMD was established by the Taiwan Environmental Protection Administration Executive Yuan in 1998 to record daily concentrations of CO and NO2 from 74 ambient air quality-monitoring stations over Taiwan Island. However, the PM2.5 data were not systematically collected in all 74 stations until 2012. Therefore, we did  not analyze PM2.5 effects on AMD in the present study. To analyze the effect of air pollution on health, we needed to link the insurant living area to the air quality-monitoring station. Because all traceable codes in the health insurance data were removed, there was no direct way to know the subject’s living area. Therefore, we used the follow strategy to predict subject’s living area: the living area for an insured person was defined based on his/her sought treatment for acute upper respiratory tract infection (AURTI) (ICD-9-CM code 460).

Study subject, exposure measurement, and comorbidity

We selected people living in areas where there were air quality-monitoring stations. Since AMD is specifically for the aged population, we only analyzed the study subjects aged 50 years or older upon enrollment. The definition of AMD used in the present study was based on the ICD-9 code (362.51 or 362.52) plus at least one of the following four procedures: color photo picture for the fundus (procedure code 23502), optical coherence tomography (OCT) exam (procedure code 23506), fluorescein angiography (procedure code 23505) or intravitreal injection (procedure code 86201). Although the dataset included medical records from the start of 1996, the data in the first few years were incomplete. Accordingly, we only analyzed the longitudinal data between the start of 2000 and the end of 2010. The patients with AMD in the present study had newly diagnosed AMD during this 11-year follow-up. People without AMD development were followed to the date of withdrawal from the program or the end of 2010.

Due to the seasonal variation of air pollution, we first calculated the average daily exposure (total exposure/days of exposure). Then we multiplied the average daily exposure with 365.25 to get the totally annual exposure since 1998 to the end of the observation year. Air pollutant concentrations were grouped into four levels based on quartile: NO2 concentration (Q1: <6563.2, Q2: 6563.2–8238.2, Q3: 8238.3–9825.5, and Q4: >9825.5 ppb), CO concentration (Q1: <195.7, Q2: 195.7–241.7, Q3: 241.8–297.1, and Q4: >297.1 ppm). Comorbidities analyzed in the present study included diabetes mellitus (DM, ICD-9-CM code 250), ischemic heart disease (IHD, ICD-9-CM codes 410–414), hypertension (ICD-9-CM codes 401–405), chronic obstructive pulmonary disease (COPD, ICD-9-CM codes 490–496), and hyperlipidemia (ICD-9-CM code 272).

Statistical analysis

The χ2 test was used to test the difference of distribution in sex, insurance fee (<14,400, 14,400–18,300, 18,301–21,000, and >21,000 New Taiwan Dollar), urbanization, and comorbidity among air pollutant quartiles. One-way analysis of variance test was used to test the mean age in each air pollutant quartile. The incidence of AMD (per 10,000 person-years) was counted in different air pollutant quartiles. The Cox proportional hazard regression was used to estimate the HRs and 95% CIs for AMD in Q2–Q4 levels of air pollution compared with AMD in the lowest one (Q1). The multivariable model was used to adjust for age, sex, insurance fee, urbanization, alcoholism, IHD, COPD, DM, hyperlipidemia, and hypertension. Smoking is a risk factor for AMD.16 Since the dataset did not provide information for the smoking status, alcoholism and COPD were used as surrogates for smoking in the multivariable analysis, which were used in our previous studies.7 17 The Kaplan-Meier analysis was used to plot the AMD-free rate curve and log-rank test was used to test the difference among air pollution quartiles. All analyses were performed using SAS V.9.3 and the SPSS V.15.1). Data were presented as mean ±SD. All statistical tests were considered statistically significant when two-tailed p values were <0.05.


Demographic data

The present study included the follow-up data on 39,819 residents who were aged 50 or older and did not have AMD upon enrollment (table 1). After the 11-year follow-up, 1442 participants developed AMD. The mean age upon enrollment was 62.33±8.87 years, and men accounted for 47.0% of all participants. The mean follow-up period was 10.77±2.63 years for all of the participants. Most of them lived in either highly (30.1%) or moderately (32.5%) urbanized areas.

Table 1

Baseline demographics in participants upon enrollment

The demographic information separated by the four levels of ambient NO2 exposure is shown in table 2. The residents in the lowest NO2 exposure areas were oldest but had the lowest incidence of AMD (2.8%), while the residents in the highest NO2 exposure areas had the highest incidence AMD (5.4%, table 2).

Table 2

Baseline characteristics according to quartiles of nitrogen dioxide levels

The demographic data according to the CO exposure levels showed a similar trend (table 3). The highest rate (5.8%) of newly diagnosed AMD was among the people living in the area with the highest CO exposure. The mean of annually total pollution for each quantile and the incidence of AMD over the study years are shown in figure 1.

Table 3

Baseline characteristics according to quartiles of CO levels

Figure 1

The mean of annually total exposure of air pollutants in each quartile and AMD incidence across study years. AMD, age-related macular degeneration; CO, carbon monoxide; NO2, nitrogen dioxide.

Multivariable analysis for AMD risk factors

To minimize the confounding effects, we adjusted for age, sex, social status (insurance fee and urbanization), and comorbidities in the multivariable analysis. Compared with the Q1 NO2 exposure, the highest exposure (Q4) increased the AMD incidence from 25.60 to 51.52 per 10,000 person-years (table 4). Similarly, the Q4 CO level increased the incidence from 28.89 to 56.24 per 10,000 person-years. After adjusting for covariates, the adjusted HR was 1.91 (95% CI 1.64 to 2.23, p<0.001) in the Q4 NO2 exposure compared with the Q1 group (table 4). The adjusted HRs were not significant for the Q2 or Q3 NO2 exposure groups: 1.15 (95% CI 0.98 to 1.35, p=0.099) and 1.09 (95% CI 0.93 to 1.29, p=0.300), respectively (see online supplementary table).

Supplemental material

Table 4

Comparisons of AMD incidences and HRs among four levels of air pollutants

Regarding the CO levels, the adjusted HR was 1.84 (95% CI 1.57 to 2.15, p<0.001) in Q4 compared with Q1 (table 4). For the Q2 and Q3 exposure groups, the adjusted HRs were 0.74 (95% CI 0.63 to 0.88, p=0.001) and 1.00 (95% CI 0.85 to 1.17, p=0.999), respectively. The trend tests showed the significant results for both NO2 and CO (p<0.001).

Furthermore, we pooled Q1, Q2, and Q3 as a single reference group to evaluate the risk of AMD in Q4. The unadjusted HRs (ie, incidence rate ratio) in Q4 compared with the reference group in NO2 and CO were 1.83 and 2.07, and the adjusted HRs of AMD were 1.77 (95% CI 1.58 to 1.98) and 2.01 (95% CI 1.79 to 2.25), respectively.

Figure 2 and figure 3 show the Kaplan–Meier plots for the AMD-free rate according to pollutant levels. During the follow-up of 12 years, the AMD-free rate in highest level of pollutants decreased faster than Q1, Q2 and Q3 for both NO2 and CO exposure.

Figure 2

The Kaplan-Meier plot shows the AMD-free rates among quartiles of NO2 exposure. AMD, age-related macular degeneration; NO2, nitrogen dioxide.

Figure 3

The Kaplan-Meier plot shows the AMD-free rates among quartiles of CO exposure. AMD, age-related macular degeneration; CO, carbon monoxide.


This is the first report using the longitudinal human data to indicate that NO2 and CO possess potential detrimental effects on AMD development. In this large population-based study, we found that long-term exposure to the highest quartile of NO2 significantly increased the risk for AMD by almost twofold even after adjusting for potential confounding factors. Similarly, exposure to the highest quartile of CO also increased the risk for AMD by 84%. NO2 is one component of reactive nitrogen species (RNS). RNS can act together with reactive oxygen species (ROS) to damage cells. CO has long been known to be toxic for health. However, CO effects on eye diseases have been almost ignored in epidemiological studies.18 The present study implied the possible harmful consequence to the eyes after exposure to CO and NO2.

We previously reported that CO and NO2 were associated with dementia based on the same datasets. Compared with lowest exposure group, the adjusted HR for dementia was 1.5 for the highest quartile of NO2 exposure, and was 1.61 for the highest quartile of CO exposure. Based on the HRs, our data indicated that these two pollutants have stronger detrimental effects on AMD than dementia. Similar to dementia, AMD is a neurodegenerative disease which is primarily affected in the central region of the retina (ie, macula). The onset of AMD has been linked to several risk factors including chronic oxidative stress and inflammation. The present study further suggests RNS might be a novel factor contributing to AMD development.

In tables 2 and 3, the residents exposed to highest level of pollutants (Q4) had lower prevalence of IHD and COPD than those in Q1. This result may be explained as follows: the residents in Q4 had a higher income than those in Q1 (per cent of highest insurance fees 25.4% in Q4 vs 19.4% in Q1). In addition, the residents in Q4 more likely lived in high urbanization areas. Accordingly to a published report, a higher prevalence of smokers was among the low income Taiwanese.19 Therefore, lower COPD and IHD rate in Q4 is probably due to a lower smoking rate.

NO2 is a marker of traffic-related air pollution. After inhalation, NO2 slowly hydrolyzes to nitrous and nitric acid which causes inflammation resulting from lipid peroxidation and oxidative stress. Due to the inhalation route, NO2 studies have been mainly focused on the respiratory tract. However, recent studies have linked NO2 pollution to the cardiovascular and neurological systems.20 Several studies have reported that high NO2 is associated with various brain diseases including low cognitive function  and a lower functional integration in children, Parkinson disease, stroke, and dementia.7 21–24 Notably, the retina is also a part of central nervous system which is biologically reasonable to be vulnerable to NO2 intoxication.

To our knowledge, air pollution has not been reported to be a risk factor for AMD. Our study first ever used a large longitudinal cohort to demonstrate significant association between AMD and high levels of ambient NO2 and CO. However, certain limitations exist in the present study: (1) the dataset does not contain information for smoking, genetic variants, and inflammatory status which is important risk factors for AMD, (2) the assignment for the residential area was based on the locations of the clinics and hospitals where a participant sought for the treatment of acute upper respiratory infections (AURI). This approach could have underestimated the pollutant risk, because a participant exposed to the lowest level of pollutant may not have any medical record about AURI. If so, we would not be able to include such a participant to the baseline data, which resulted in an increased AMD incidence rate in the lowest exposure group. However, the overall findings are intriguing and warranted for further exploration.


In conclusion, we linked the nation health database and air quality database to report NO2 and CO as risk factors for AMD. Our results indicate the highest quartile of each pollutant could increase the risk of AMD by almost twofold, while the moderate exposure did not contribute to AMD development.



  • Contributors Conceptualization: K-HC, P-YH, S-HHJ, and C-LL; methodology: K-HC, C-JL, S-HHJ, and C-LL; formal analysis: K-HC and C-LL; investigation: all authors; writing (original draft preparation): K-HC; writing (review and editing): S-HHJ and C-LL; visualization: all authors; supervision: S-HHJ and C-LL; project administration: S-HHJ and C-LL.

  • Funding This study was supported by the grants from the Asia University-CMU Hospital intramural grant (ASIA-106-CMUH-18), Tungs' Taichung Metroharbor Hospital (TTMHH-108R0029), Taiwan Ministry of Health and Welfare Clinical Trial Center (MOHW108-TDU-B-212-133004), Taiwan Ministry of Science and Technology (MOST 107-2321-B-039-004 and MOST 106-3114-8-039-002), the “Drug Development Center, China Medical University" from The Featured Areas Research Center Program within the framework of the Higher Education Sprout Project by the Ministry of Education in Taiwan.

  • Competing interests None declared.

  • Ethics approval The study was approved by the Ethics Review Board of China Medical University and Hospital, Taiwan (CMUH-104-REC2-115).

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Data sharing statement Data are available from the NHIRD published by Taiwan National Health Insurance Bureau. Due to the ‘Personal Information Protection Act’, data cannot be made publicly available (

  • Patient consent for publication Not required.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.