Air Pollution and Chronic Eye Disease in Adults: A Scoping Review

ABSTRACT

Purpose

We conducted a scoping review of studies examining ambient air pollution as a risk factor for chronic eye disease influencing the lens, retina, and intraocular pressure in adults.

Methods

Terms related to air pollution and eye disease outcomes were used to search for publications on Embase, Web of Science Core Collection, Global Health, PubMed, and the Cochrane Central Register of Controlled Trials from January 1, 2010, through April 11, 2022.

Results

We identified 27 articles, focusing on the following non-mutually exclusive outcomes: cataract (n = 9), presbyopia (n = 1), retinal vein occlusion or central retinal arteriolar and venular equivalents (n = 5), intraocular pressure (IOP) (n = 3), glaucoma (n = 5), age-related macular degeneration (AMD) (n = 5), diabetic retinopathy (n = 2), and measures of retinal morphology (n = 3). Study designs included cross-sectional (n = 16), case–control (n = 4), and longitudinal (n = 7). Air pollutants were measured in 50% and 95% of the studies on lens and retina or IOP, respectively, and these exposures were assigned to geographic locations. Most research was conducted in global regions with high exposure to air pollution. Consistent associations suggested a possibly increased risk of cataract and retina-associated chronic eye disease with increasing exposure to particulate matter (PM_2.5-PM₁₀), NO₂, NO_x, and SO₂. Associations with O₃ were less consistent.

Conclusions

Accumulating research suggests air pollution may be a modifiable risk factor for chronic eye diseases of the lens and retina. The number of studies on each specific lens- or retina-related outcome is limited. Guidelines regarding the role of air pollution in chronic eye disease do not exist.

KEYWORDS:

• Epidemiology
• air pollution
• particulate matter
• retina
• cornea
• lens
• reveiw
• systematic review
• environmental exposure

Introduction

Approximately 43.3 million people are blind,¹ 295 million have moderate and severe vision impairment, and 258 million have mild vision impairment.¹ Between 1990 and 2020, the prevalence of moderate and severe vision loss and the overall burden of blindness and vision impairment in individuals older than 50 has increased.¹ Significant causes of blindness in this age group in descending order include cataracts, glaucoma, undercorrected refractive error, age-related macular degeneration (AMD), and diabetic retinopathy.² Individuals with vision impairment and blindness are at an increased risk of falls and depression,³ decreased productivity, and increased healthcare costs.⁴

Researchers are exploring the impacts of ambient air pollution on numerous disease outcomes. Most publications on this topic have focused on associations between ambient air pollution exposure and corneal outcomes, such as dry eye disease. Yet, we are only beginning to understand how air pollutants may contribute to chronic eye disease beyond those that influence the cornea in direct contact with the environment. Whereas the lens, retina, and ocular nerve are not directly exposed to air pollution, these tissues may be affected through increased circulating levels of inflammatory biomarkers and reactive oxygen species produced in response to particulate matter (PM) exposure, for example (reviewed in).⁵ Our objective was to summarize the scholarly publications on the relationship between air pollution and chronic eye disease of the lens and retina, as well as intraocular pressure (IOP), a known risk factor for glaucoma, and measures of retinal layer thickness as indicators of AMD and glaucoma.

Methods/literature search

Search terms were derived with the use of PubMed’s medical subject headings [MeSH]⁶ and the Yale MeSH analyzer tool.⁷ Searches, limited to the English language only and excluding animal studies, were conducted using Embase, Web of Science Core Collection (WoS), Global Health, PubMed, and Cochrane Central Register of Controlled Trials on publications from January 1, 2010, through April 11, 2022. Table 1 contains search terms used in WoS.

Table 1. Web of Science search on 01–09–2020.

TOPIC: ((Air Pollutant* OR Particulate Matter OR Air Pollution OR Nitrogen Dioxide OR Nitrogen Oxide OR Vehicle Emission* OR Traffic-Related Pollution OR Ozone OR Sulfur Dioxide OR Sulfur Oxide* OR Inhalation Exposure OR Carcinogen*, Environmental OR Carbon Monoxide OR Smog OR Environmental Exposure)) AND TOPIC: ((Macular Degeneration OR Conjunctivitis OR Glaucoma OR Eye Disease* OR Dry Eye Syndrome* OR Xerophthalmia OR Blepharitis OR Cataract* OR Conjunctiva OR Retinal Disease* OR Retinal Vessel* OR Corneal Disease* OR Retinal Degeneration OR Retinal Drusen OR Optic Disk Drusen OR Optic Nerve Disease*OR Retina*OR Cornea* OR Optic Nerve* OR Retinoblastoma*))

Refined by: LANGUAGES: (ENGLISH)

In Phase I, two independent reviewers screened the references on the title and abstract (AEM and SD). References were excluded if it was evident that they were: 1) not written in English, 2) a duplicate of another paper, 3) animal or cell culture studies, 4) not conducted in adults (≥18 years), 5) case reports, 6) editorials, letters, or replies, 7) chapters, review articles, or meta-analyses, or 8) conference papers and abstracts. Also excluded were papers focusing on mustard gas, chromium, lead, mercury, heavy metals, organic solvents, pollen, ultraviolet light or radiation, ocular cancer, conjunctivitis, ocular inflammation, ocular allergy, or Xerophthalmia. In the case of discrepancies, both individuals re-reviewed the titles and abstracts and made a final joint decision regarding whether to include the reference. In Phase II, the entire paper for each reference identified from Phase I was reviewed by the two reviewers independently for inclusion and exclusion using the criteria outlined above. The reviewers discussed all papers with classification discrepancies and jointly resolved whether to include each paper. Further refinement of exclusion criteria was extended at the end of Phase II to exclude manuscripts with ecologic study designs with aggregated outcome measures as well as papers focusing on carbon monoxide poisoning, uveitis, pterygium, or general vision impairment. For each exclusion, a specific reason was recorded. Papers on air pollution and the cornea were identified, although they are not described in this review.

The papers selected with a focus on lens, retina, and IOP were entered into review tables. Data was collected on the author, publication year, overall sample size, overall case number, study design, the period during which the study was conducted, information on the exposure measure used, information on how the outcome was assessed, whether covariates were adjusted for in regression models, and the magnitude and direction of association for each exposure measured in relation to each outcome assessed. Data regarding the detailed search terms used for all search engines, endnote files, or reasons for study exclusion in Phase II are available upon request from the corresponding author.

Results

The search recovered 5,136 results from Embase, 794 from WoS, 93 from Global Health, 1,807 from PubMed, and 26 from Cochrane, resulting in 7,856 references. In Endnote, 981 duplicates were eliminated, leaving 6,875 unique references (Figure 1). In Phase I, the reviewers identified 131 unique references for full paper review. In Phase II, the reviewers included 63 full papers. Papers focused exclusively on the lens (n = 8, Supplemental Table 1), retina, or intraocular pressure (n = 17, Supplemental Table 2). Two papers were identified that included outcomes of both the lens and retina (Supplemental Tables 1 & 2). We identified 26 papers exclusively on cornea but did not summarize them; however, we provide the identified references for the reader.^8–43 All discussed studies in this review adjusted for potential confounders and risk estimates presented are from adjusted models.

Figure 1. Flow Diagram for the Scoping Review on Air Pollution and Chronic Eye Diseases in Adults.

Lens

Cataract

Nine studies^44–52 that examined cataract were identified, of which five were cross-sectional,^{45–47,51,52} one was a case–control study,⁴⁴ and three were longitudinal.^48–50 Data from the Korea National Health and Nutrition Examination Survey (KHANES) was used to examine associations between PM₁₀, ozone (O₃), nitrogen dioxide (NO₂), and sulfur dioxide (SO₂), and graded slit-lamp photographs for any cataract and cataract subtype (n = 19,910).⁴⁷ Contrary to their hypothesis, researchers observed protective associations between O₃ and any cataract and nuclear cataract (NC) and protective associations between NO₂ and both NC and anterior subcapsular (ASC) cataract. The odds ratio (OR) and (95% confidence intervals [CI]) for any cataract and NC per 0.003 ppm unit increase of O₃ were 0.80 (0.69–0.93) and 0.73 (0.61–0.86), respectively. The ORs (95% CIs) were 0.69 (0.52–0.93) and 0.88 (0.79–0.97) for any ASC and NC, respectively, per 0.003 ppm unit increase of NO₂. In a retrospective cohort study (n = 115,728), also conducted in Korea,⁴⁸ objectively measured air pollution estimates (PM₁₀, PM_2.5, NO₂, carbon dioxide [CO], SO₂, and O₃) were examined in relation to claims data for newly developed cataract, defined as a first diagnosis, and cataract surgery during the follow-up period. They observed an increased risk for incident cataract among participants in quartile four compared to one for concentrations of PM₁₀, NO₂, and SO_2. The hazard ratio (HR) (95% CIs) was 1.069 (1.02–1.15) for exposure to PM₁₀, and the risk of cataract was of a similar magnitude for NO₂ and SO_2. This study also observed a decreased risk of cataract (0.931 [0.888–0.977]) among participants in quartile four compared to one of O₃ concentrations. The risk of cataract was significantly associated with these exposures in individuals ≥65 years, but not <65, in women but not men, except for PM_10, which was still associated with increased risk in men.

Only two studies used objective measures of air pollution. They are also the cataract studies conducted outside Asia. One study was a cross-sectional analysis of the Canadian Longitudinal Study on Aging,⁵¹ and the other was a prospective study on participants of the United Kingdom (UK) Biobank.⁵⁰ In the Canadian Study, increasing concentrations of O₃ (ppb) were associated with a decrease odds of self-reported cataract (OR [95% CI] = 0.92 [0.85–0.99]). In the UK Biobank cohort (n = 433,727), measures of PM_2.5, PM_2.5–10, PM₁₀, PM_2.5 absorbance (a proxy for black carbon), NO₂, and nitrogen oxides (NO_x) were associated with the increased risk of incident cataract surgery assessed using the National Health Service procedures statistics. The authors observed a significant increased risk of surgery in participants in quartile four versus one for PM_2.5 (HR [95% CI] = 1.14 [1.08–1.19]), NO₂ (1.11 [1.06–1.17]), and NO_x (1.09 [1.04–1.15]).

The other published studies on air pollution and cataract have not included objective measures of air pollution and have focused on exposure to fuel types used in cooking. In India,⁴⁶ researchers examined associations between household cooking fuels and any cataract and cataract subtype in 5,871 participants. Current use of biomass versus clean fuel was associated with an increased odds of NC, OR (95% CI) = 1.24 (1.01–1.42) in all participants. This significant direct association was also observed only in women (1.46 [1.16–1.84]) but not men. In women but not men, positive associations were observed between years of biomass fuel use and any cataract (1.18 [1.02–1.36]) and NC (1.28 [1.10–1.48]), and between the use of any kerosene (yes/no) and NC (1.75 [1.04–2.97]) and PSC (1.71 [1.10–2.64]). Researchers in Nepal [46] use slit-lamp photographs to identify cataract subtype and nuclear color. The odds of cataract were examined in relation to the type of cooking stove used (n = 143 women). Higher odds of nuclear opacity were observed with the use of biomass versus gas fuel (OR [95% CI] = 2.58 [1.22–5.46]) and duration of biomass fuel use of >40 years versus no use (4.29 [1.21–15.25]); kerosene use compared to gas was associated with a higher likelihood of nuclear color (5.48 [1.23–24.37]). In a case–control study⁴⁴ in Bangladesh (n = 459), women had an increased adjusted odds of cataract if they ever versus never used rice straw (1.95 [1.03–3.69]) and a decreased odds if they ever versus never used cow dung for fuel (0.45 [0.24–0.84]). There was an increased odds of cataract among men using versus not using wood/dry leaves for fuel (3.51 [1.24–9.95]). This study did not investigate the subtype of cataract. A recently published cross-sectional study (n = 31,373) of participants from six low- to middle-income countries examined associations between the type of cooking fuel used and the self-reported prevalence of cataract.⁵² The study findings are consistent with results from other studies, with unclean versus clean fuel use associated with increased adjusted odds of self-reported cataract in all countries combined (OR [95% CI] = 1.42 [1.29–1.56]). The significant findings were restricted to China and India in analyses stratified by country. They also observed an increased odds of cataract (1.10 [1.10–1.21]) in individuals using solid fuels (coal or biomass fuels) in open stoves or fires compared to closed stove settings. In a large prospective study of participants in the China Kadoorie Biobank (n = 486,832),⁴⁹ the odds of incident cataract identified with a national health insurance database, death, and disease registries were increased in those identified as long-term solid fuel users versus never-regular cooks (OR [95% CI] = 1.17 [1.08–1.26]). This association remained direct but not statistically significant when these data were analyzed using available time-to-event data. A longer duration of solid fuel use, but not type, was also associated with an increased odds of cataract.

Presbyopia

Only one study examined PM_2.5 and O₃ in relation to self-reported prevalent presbyopia. Participants were from six low- to middle-income countries,⁵³ and direct associations of presbyopia and both PM_2.5 and O₃ (per 10 µg/m³ increase) were detected: 1.15 (1.09–1.21) and 1.37 (1.23–1.54), respectively. A higher prevalence of presbyopia was not observed until PM_2.5, and O₃ concentrations reached 15 µg/m³ and 55 µg/m³, respectively. A synergistic interaction was observed with a significant two-fold increased odds of presbyopia in those categorized as high compared to low for both pollutants.

Retina

Retinal microvascular equivalents

Until recently, most studies on retinal diseases and air pollution have examined central retinal arteriolar or venular equivalents (CRAE and CRVE, respectively).^54–58 All have been cross-sectional except for one recent retrospective cohort study.⁵⁸ The first paper analyzed data from the Multi-ethnic Study of Atherosclerosis.⁵⁴ Adar et al. found an inverse, cross-sectional association between estimates of long-term (years) PM_2.5 concentrations and CRAE equal to a decrease in 0.8 µm in vessel width per interquartile range (IQR) of PM_2.5 in µg/m³. No statistically significant association was found with short-term (days before exam) PM_2.5 and CRAE or between short- or long-term PM_2.5 measures and CRVE. Next, three small (n < 100 subjects), cross-sectional studies were conducted by a research team in Belgium.^55–57 Each study assessed recent black carbon (BC) and/or PM₁₀ measures and CRAE or CRVE hours to days later. PM₁₀ was inversely associated with CRAE.^55,57 A decrease of 0.93⁵⁵ and 0.72⁵⁷ µm in CRAE was observed with each 10 µg/m³ increase in PM₁₀ (assessed in the previous 24 hours). Inconsistent results were observed with the association between PM₁₀ and CRVE.^55,57 Two studies measured both short (preceding 48 hours) and sub-chronic (preceding 7 days) BC exposure.^55,56 One study found that short-term BC exposure was inversely associated with CRAE (β-coefficient [95% CI] = −1.84 [−3.18 to −0.51] µm decrease per 1 µg/m³ increase in BC).⁵⁵ The other found that sub-chronic BC exposure was directly associated with CRVE (β-coefficient [95% CI] = 5.65 [1.33 to 9.96] µm increase per 0.631 µg/m³ increase in BC).⁵⁶ Different from microvascular equivalents, a large retrospective study using ICD-9 codes for retinal vein occlusion (RVO) observed an increased risk of RVO with higher concentrations of total hydrocarbons (~20-fold increased risk per 0.51-ppm increase) and nonmethane hydrocarbons (~4-fold increased risk per 0.27-ppm).⁵⁸

Intraocular pressure and glaucoma

Three studies^51,59,60 examined associations with intraocular pressure (IOP), a risk factor for glaucoma; two were cross-sectional,^51,59 and one was longitudinal in design.⁶⁰ In a large, cross-sectional community-based cohort in the UK,⁵⁹ clinical measurements of IOP were directly associated with higher residential annual average PM_2.5. Corneal-compensated and Goldmann-corrected IOP increased 0.03 and 0.04 mm Hg, respectively, for each IQR increase in PM_2.5. Differently, repeated measures from the US Department of Veterans Affairs Normative Aging Study⁶⁰ observed no statistically significant associations between BC exposure estimates over a year and average IOP measures. However, they did observe a direct association between BC and IOP (β-coefficient [95% CI] = 0.36 [0.003–0.73] mm Hg increase per IQR increase in BC) in participants with a high allelic score for genes involved in oxidative stress metabolism. In the previously mentioned Canadian Longitudinal Study on Aging,⁵¹ an increase in one IQR of PM_2.5 was associated with a 0.24 mm Hg increase in IOP, similar to the BC effect in the Normative Aging Study.

Five studies examined associations with glaucoma^{49,51,61–63}; two cross-sectional^51,63 and three longitudinal.^49,61,62 In the previously mentioned China Kadoorie Biobank (n = 486,832) prospective study,⁴⁹ no statistically significant associations were observed between the odds of glaucoma and long-term solid fuel use, duration, or type of solid fuel used. However, one cross-sectional study in China⁶³ and two nested case–control studies in Taiwan^61,62 observed high compared to low PM_2.5 exposure associated with a greater odds or risk of glaucoma. Yang et al.⁶³ observed increased odds of primary angle-closure glaucoma (PACG) (OR [95% CI] = 1.14 [1.02–1.26]) for each 10 µg/m³ increase in PM₁₀. No significant observations were observed with primary open-angle glaucoma (POAG) or a combined glaucoma endpoint and PM_2.5. Two studies used claims data from Taiwan’s National Health Insurance Research Database (NHIRD) to identify glaucoma.^61,62 One study observed a 1.668-fold increased odds of incident POAG for individuals with ≥2 times the reference value of PM_2.5 (<25 µg/m³).⁶² A similar magnitude of increased odds of glaucoma among patients in quartile 4 compared to 1 for PM_2.5.was observed among a sample of individuals with diabetes within the NHIRD.⁶¹

Age-related macular degeneration

Five papers on AMD have been published, three are cross-sectional^51,64,65 and two longitudinal.^66,67 The first paper on air pollution and AMD⁶⁶ was a prospective analysis using Taiwan’s Longitudinal Health insurance database (n = 39,819). They observed a higher incidence of AMD, diagnosed by ICD-9 codes, in those with high compared to low concentrations of NO₂ (hazard ratio [HR] [95% CI] = 1.91 [1.64–2.23], p for trend < 0.001) and CO (1.84 [1.5–2.15], p < 0.001). The most recent study on air pollution and AMD was also longitudinal.⁶⁷ It involved a large retrospective population-based cohort of persons in Taiwan with data on incident AMD claims from the Taiwan National Health Insurance Research Database linked to geographically defined estimates of daily PM_2.5 exposure from 2001 to 2011. The HR (95% CI) for incident AMD per 10 µg/m³ increase in PM_2.5 was 1.19 (1.13–1.25).

The cross-sectional analysis using data from the Canadian Longitudinal Study on Aging⁵¹ reported an increased odds of self-reported visually impairing AMD with an increase of one IQR in PM_2.5 (OR [95% CI] = 1.5 [1.10–20.9]). Similarly, a study with participants of the UK Biobank observed an increased odds of self-reported AMD with continuous increasing concentrations (µg/m³) of PM_2.5.⁶⁵ Ju et al.,⁶⁴ using data from KNANES, observed an increased odds (~20%) of any and early AMD with each IQR change in NO₂ or CO but a decreased odds (~20%) of any and early AMD with each IQR change in O_3. For these observations, the air pollution estimates were assigned to the same year as KNHANES assessments of AMD. Additionally, using exposure estimates from earlier time points, the authors found increased odds of early AMD with exposure to PM₁₀ in the prior 3–6 years. When the exposure data was parameterized as high versus low rather than as continuous, the PM₁₀ association remained and existed for all years of exposure (current and prior 1–5 years); increased odds of early AMD was observed for high versus low exposure to SO₂ in the prior 1–2 years.

Diabetic retinopathy

Two studies to examine associations with diabetic retinopathy.^68,69 Shan et al.⁶⁸ examined clinically diagnosed diabetic retinopathy with long-term exposure estimates to PM_2.5 in a nationally representative sample of Chinese adults with diabetes. They observed 1.4-fold increased odds of retinopathy (OR [95% CI] = 1.41 [1.27–1.57]) for a 10 µg/m³ increase in PM_2.5. In the second study, Pan et al.⁶⁹ conducted a nested case–control study using data from individuals enrolled in the National Health Insurance Program in Taiwan. Individuals with diabetes were considered to have diabetic retinopathy if they had claims data indicative of this eye disease one year or later after their diagnosis of diabetes. The odds of diabetic retinopathy in relation to previously collected yearly average concentrations of PM_2.5, PM_2.5–10, PM₁₀, CO, NO₂, SO₂, and O₃ were assessed. Increasing concentrations of 10 µg/m³ of PM_2.5, PM_2.5–10, and PM₁₀ were directly associated with increased odds of diabetic retinopathy with ORs of 1.29, 1.37, and 1.18, respectively.

Retinal morphology

Three cross-sectional studies in the UK Biobank cohort examined air pollution estimates to morphologic measures in the retina assessed with spectral domain ocular coherence tomography (SD-OCT). Ganglion cell-inner plexiform layer (GC-IPL) thickness was inversely associated with PM_2.5 concentrations.⁵⁹ An increase in one IQR of PM _2.5, PM_2.5 absorbance, and NO_x in µg/m³ was associated with decreases of 0.10 to 0.16 µm in the thickness of the photoreceptor synaptic region.⁶⁵ Increases of one IQR for PM_2.5, PM_2.5 absorbance, PM₁₀, NO₂, and NO_x were associated with increases in either photoreceptor inner and outer segments (0.3–0.12 µm), and an increase in one IQR of NO₂ was associated with a 0.15 µm increase in the total length of the photoreceptor.⁶⁵ And more inverse associations were observed between exposure to air pollution measures and the thickness of inner than outer retinal layers.⁷⁰

Discussion

The interest in the association between air pollution and chronic eye disease has increased in the last few years, with 13 (48%) of the identified 27 papers published in 2021 and 2022. Of the papers reviewed, the greatest number of studies focused on cataract, followed by retinal microvasculature, glaucoma, and AMD, with diabetic retinopathy and presbyopia being the least studied.

Concerning studies of the lens, only two recently conducted studies were longitudinal with incident outcomes.^49,50 One used ICD-10 codes to identify incident cases,⁴⁹ the other identified incident cataract surgery cases from medical databases,⁵⁰ and neither study examined cataract subtype. Only the prospective study using incident cataract surgery⁵⁰ used objective air pollution measures. They observed positive associations between PM_2.5, NO₂, and NO_x and incident cataract surgery. The other prospective study⁴⁹ examined fuel use in cooking as a risk factor for first diagnosis with cataract. Like other cross-sectional and case–control studies examining fuel use and cataract,^44–46,52 solid versus clean fuel use/no cooking was associated with an increased risk of cataract; and often observed in women but not men. Four studies observed differences in associations between air pollution exposure and cataract by sex.^44,46,48,49 All of these studies, but Tanchangy et al.,⁴⁴ adjusted for sex in multivariable models and still observed positive associations between air pollution measures and cataract in the overall samples, with men and women combined, in addition to positive observations in just women. Women likely engage in more cooking activities and thus experience greater ocular exposure to cooking fuel sources. Perhaps as a group, women have greater total exposure to air pollution, than men, and this higher threshold is needed to increase risk of cataract.

Only four cataract studies used objectively measured air pollution estimates,^47,48,50,51 and all studies observed direct associations between measures of air pollutants (PM_2.5, PM₁₀ SO₂, CO, NO₂, or NO_x) and the odds of cataract. However, in all three studies^47,48,51 that examined O₃ exposure, inverse associations were observed between O₃ concentrations and cataract. The authors have hypothesized that ozone may be protective against cataracts because it absorbs ultraviolet radiation, a known risk factor for cataract.⁷¹ Only three studies on cataract examined disease subtypes. Of those, ASC, NC, and PSC, were found to be associated with air pollution,^45–47 but not CC in the two studies it was examined.^45,46 Whether the lens cortex is as susceptible to damage from air pollution region remain to be determined.

Most studies on cataract were conducted in countries with high air pollution, including Bangladesh, India, China, or Nepal.⁷² Cataract surgery is not an accessible option in many such regions.⁷³ Only recently, researchers have examined air pollution and cataract associations in countries with less burden of air pollutants, such as the United Kingdom or Canada.^50,51 Because of this, there remains a need to understand better how the risk of cataract varies across the distribution of air pollution levels globally, inclusive of lower levels of exposure. And only one study has focused on presbyopia showing a negative association with higher versus lower PM_2.5 and O₃ concentrations.⁵³

Most studies on retinal disease have measured air particulate matter (PM_2.5, PM₁₀) and BC and found exposure to be directly associated with adverse retinal microvasculature outcomes. Air pollution is thought to lead to the narrowing of microvascular arterials⁷⁴ and decreased availability of NO.⁷⁵ An alternative mechanism is that air pollutants trigger sympathetic nerve activity causing retinal vessel smooth muscle constriction.⁷⁶ Only one study examined RVO, and they observed that TCH and NMHC were associated with incident diagnosis of RVO. Evidence that air pollution affects the health of the microvascular as well as macrovascular, is building. Retinal vasculature can be examined non-invasively and often reflects the function of the microvasculature elsewhere, like the kidney.⁷⁷ However, three of five published studies on retinal microvasculature were limited by small sample sizes.^55–57 Studying changes in microvascular width could be challenging and may need repeated measures over years to truly understand patterns of air pollution on microvasculature response. Evidence to date focuses on air pollution measures followed by hours to days later for CRAE and CRVE, providing temporality but an inability to model change in CRAE or CRVE over long periods of time,^55–57 and the extent to which that is important is unknown.

The hypothesis that air pollution could cause neuroinflammation, including the optic nerve,⁷⁸ has stimulated research on air pollution and glaucoma. All studies on IOP supported evidence that PM_2.5, BC, and O₃ are understudied risk factors for glaucoma.^51,59,60 Five studies examined glaucoma as an outcome^49,61,62 and also found direct associations with PM_2.5. Only one other exposure, O₃, was examined in relation to glaucoma but was not associated with this outcome. Two of these studies had robust designs as nested-case control studies^61,62 but were limited as medical claims data were used to identify incident cases. Only one study⁶³ examined glaucoma subtype and observed an association with PACG, the less common form of glaucoma. Additional work by glaucoma subtype is needed. Studying glaucoma as a clinical outcome is challenging as measuring functional visual loss requires visual field testing.⁷⁹ Unlike other eye diseases, it may be essential to measure this outcome clinically to obtain valid data.

A robust study examined incident AMD with NO₂ and CO and found a direct association.⁶⁶ Four additional papers published in the last 2 years support this original observation.^51,64,65,67 Three focused only on PM_2.5.^51,65,67 The fourth study observed cross-sectional associations with CO, NO₂, PM₁₀, and SO₂, but an inverse association with O_3, similar to the protective association seen with O₃ and cataract. Ultraviolet light is also a risk factor for AMD⁸⁰ and perhaps the protective effect of O₃ on ultraviolet light also influences the effect of O₃ on AMD. Two papers also support direct associations between PM subtypes and diabetic retinopathy.^68,69 Limitations of these studies include the use of self-reported AMD status^51,65 or claims data for AMD⁶⁷ or diabetic retinopathy.⁶⁹ No studies have examined the progression of AMD or diabetic retinopathy, in addition to incidence, over time.

Recent findings from the UK Biobank study found that air pollution was associated with OCT-assessed measures of retinal layers, such as the retinal nerve fiber layers and GC-IPL. However, the direction of associations between retinal nerve fiber layer and GC-IPL thickness and AMD and glaucoma risk is still to be determined.⁷⁰ Continued study of OCT measures of the neurosensory retina and air pollution are needed for glaucoma and AMD. Concerning AMD, it would allow for measurements of AMD subtypes, such as pseudodrusen, that cannot be identified from fundus photographs.⁸¹

Our study was a scoping review and did not consider ocular outcomes related to the cornea (e.g., dry eye disease) or others, such as conjunctivitis or ocular cancer. We also did not conduct a meta-analysis to summarize effect estimates across studies. All studies with objective air pollution measures were still limited by the assessment of ecologic measures of air pollution. No studies assessed exposure by having participants carry portable air pollution monitors. This would help tease out confounding by neighborhood socioeconomic status. Furthermore, all studies reviewed are observational and subject to the possibility of residual confounding. For all ocular outcomes, data is lacking on associations across the full range of air pollution levels, not just in areas with high exposure. Further understanding is needed regarding interactions between ambient air pollution, lifestyle behaviors (diet and smoking), and genetics. Studies to date primarily involve observational designs for ethical reasons; however, one could imagine studies on eye outcomes when air pollution is reduced.

Conclusions

Based on the literature reviewed, increasing exposure to particulate matter (PM_2.5-PM₁₀), NO₂, NO_x, and SO₂ are associated with an increased risk of cataract and retina-associated chronic eye disease. Associations with O₃ were less consistent and could be protective against cataract or AMD. Objective measures of air pollution at the individual level and studies in geographic areas with lower levels of air pollution exposure than in Asia are lacking. The number of studies on each specific lens- or retina-related outcome is limited. Guidelines regarding the role of air pollution in chronic eye disease do not exist. We conclude that there is room for a better understanding of the influence of air pollution on retinal outcomes, especially AMD and glaucoma, both of which have limited treatment options.

Disclosure statement

There are no conflicts of interest for any authors.

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/09286586.2023.2183513

Additional information

Funding

The author(s) reported that there is no funding associatedwith the work featured in this article.