Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2021 Jun 1.
Published in final edited form as: J Glaucoma. 2020 Jun;29(6):448–455. doi: 10.1097/IJG.0000000000001496

Cohort study of non-melanoma skin cancer and the risk of exfoliation glaucoma

Jae H Kang 1, Trang VoPham 1,2, Francine Laden 1,3,4, Bernard A Rosner 1,5, Barbara Wirostko 6, Robert Ritch 7, Janey L Wiggs 8, Abrar Qureshi 9,10, Hongmei Nan 11,12, Louis R Pasquale 1,7,13
PMCID: PMC7317065  NIHMSID: NIHMS1576595  PMID: 32487970

Abstract

Purpose:

To evaluate the relationship between non-melanoma skin cancer (a marker of ultraviolet radiation exposure) and exfoliation glaucoma (XFG).

Methods:

We performed a cohort study of US women (n=79,102; 1980–2014) and men (n=41,205; 1986–2014), aged 40+ years and at risk for glaucoma who reported eye exams. From 1984 (women)/1988 (men), we asked about basal cell carcinoma (BCC) or squamous cell carcinoma (SCC) history separately; in prior years, we asked about any non-melanoma skin cancer history in a single question. SCC was confirmed with histopathology reports while BCC and any early (<1984/<1988) non-melanoma skin cancer history was self-reported. Incident XFG cases (362 women and 83 men) were confirmed with medical records. Using pooled data, we estimated multivariable-adjusted relative risks (MVRR; 95% confidence intervals [CIs]) with Cox proportional hazards models that were stratified by age (in months), 2-year time period at risk and average lifetime residential latitude.

Results:

In multivariable-adjusted analyses, we observed a 40% higher XFG risk with any non-melanoma skin cancer history (MVRR=1.40; 95% CI=1.08,1.82); the association was observed even with 4 and 8 year lags in non-melanoma skin cancer history. Also, the non-melanoma skin cancer association was stronger in younger (<65 years; MVRR=2.56; 95% CI=1.62,4.05) versus older participants (≥65 years; MVRR=1.25; 95% CI=0.94,1.66; p for interaction=0.01) and those living in northern latitudes (≥42° north; MVRR=1.92; 95% CI=1.28,2.88) versus more southern latitudes (<42° north; MVRR=1.19; 95% CI=0.86,1.66; p for interaction=0.04).

Conclusions:

Non-melanoma skin cancer was associated with higher XFG risk, particularly among younger participants and those living in Northern US.

Keywords: exfoliation syndrome, glaucoma, ultraviolet light, skin cancer, epidemiology

PRECIS

In a cohort study of 120,307 participants with 25+ years of follow-up, a history of non-melanoma skin cancer was associated with a 40% higher exfoliation glaucoma risk.

INTRODUCTION

Exfoliation syndrome (XFS) is the most common identifiable cause of open-angle glaucoma.1 XFS has been associated with considerable ocular morbidity including glaucoma [exfoliation glaucoma (XFG)],2 keratopathies, cataract,3 intraocular lens subluxation4 and retinal venous occlusive disease.5–7 Exfoliation material (XFM), combined with pigment from the iris sphincter region, blocks the trabecular meshwork, leading to elevated intraocular pressure (IOP) and glaucoma.8 While the discovery of LOXL1,9 POMP10 and other genetic loci11 has advanced our understanding, the etiology of XFS remains largely unknown.

Evidence suggests that greater ocular UV radiation exposure may be associated with increased XFS/XFG risk. While ambient UV radiation and UV radiation to the skin is greatest near equatorial regions, ocular UV exposure may be greatest in latitudes farthest from the equator; indeed, living farther from the equator is an established risk factor for XFS and XFG.12–19 Areas close to the polar regions are characterized by higher UV intensity20,21 and larger amounts of snow, which can reflect 80% of the sun’s rays22 may increase the intensity of ocular UV exposure. In a case-control study,23 a history of work over water or snow was associated with ~4-fold increased odds of XFS, while greater time spent wearing sunglasses when outside in the summertime was associated with lower odds. Because ocular surface UV exposure is difficult to measure objectively,24 additional studies using various measures correlated with greater ocular UV exposure are needed.

One indirect measure of greater personal ocular UV exposure may be a history of non-melanoma skin cancer (NMSC), which would be a marker of personal erythemal UV exposure.25 Previously, in a similar cohort study of younger female nurses,26 compared to the lowest quintile, the highest quintile of long-term adulthood UV flux (a measure that incorporated state-level latitude, elevation and cloud cover only27–29) was associated with a 2.35 fold higher basal cell cancer (BCC) risk (p for trend<0.0001) and a 2.53 fold higher squamous cell cancer (SCC) risk (p for trend=0.009). This is consistent with existing literature that NMSC is associated with chronic UVA and UVB exposure.30–32 Furthermore, compared with the lowest quartile of plasma 25-hydroxyvitamin D (25(OH)D), the highest quartile had a 2.07 higher BCC odds (p for trend<0.0001) and 3.77 higher SCC (p for trend=0.0002),33 a finding consistent with UVB-induced dermal synthesis of vitamin D.34 In particular, higher plasma 25(OH)D was most strongly associated with BCC in women whose blood was collected outside the summer season or who were from areas with less UVB flux, supporting the notion that NMSC represents a valid marker of greater personal UV exposure. Thus, we evaluated the relationship between a history of NMSC, which would be a marker of personal UV exposure,25 and the risk of XFG.

METHODS

Study population

We included participants from the ongoing Nurses’ Health Study (NHS; initiated in 1976 as an observational cohort study of oral contraceptives and cancer when 121,700 US registered female nurses aged 30 to 55 years completed a mailed questionnaire35) and the Health Professionals Follow-up Study (HPFS; similarly initiated as an observational study of diet and health outcomes in 1986 with 51,529 male health professionals36). Participants complete ongoing biennial questionnaires about their lifestyle and diseases, such as glaucoma. The follow-up was high (>85%). The study protocol was approved by the institutional review boards (IRBs) of Partners HealthCare and Harvard T. H. Chan School of Public Health; they ascertained that participants’ completed questionnaires served as implied consent.

The study period was from “baseline” (1980 in NHS and 1986 in HPFS) to 2014. Participants contributed person-time in approximate 2-year units with updated information derived from mailed questionnaires, starting at baseline until the earliest occurrence of glaucoma, cancer except NMSC (as cancer may cause major changes in lifestyle), death, loss to follow-up or 2014 (study end). A participant could contribute person-time of observation, only if the participant was at least 40 years of age (as glaucoma risk increases from age 40+ years) and indicated an eye exam in the 2-year risk period (to minimize detection bias).

Among the 121,700 women and 51,529 men, at baseline, participants were excluded for the following reasons: 1) 29,233 women and 1,596 men who did not complete the initial semiquantitative food frequency questionnaire or provided inadequate data (as the initiation of our studies of glaucoma in these cohorts were motivated by conducting studies of dietary risk factors); 2) 3,622 women and 1,927 men with prevalent cancer (excluding NMSC); 3) 851 women and 1,037 men with prevalent glaucoma or glaucoma suspect; 4) 766 women and 930 men who were lost to follow-up within 2 years after baseline; 5) 5,523 women and 3,195 men who did not report any eye exams during follow-up; 6) 174 women and 678 men who underwent cataract extraction (which hinders the detection of XFS), and 7) 43 women and 64 men who reported SCC that could not be confirmed with medical records. In addition, at each 2-year risk period, we included only those participants who were aged 40+ years and reported eye exams in the risk period. By 2014, a total of 79,102 women and 41,205 men contributed person-time.

Case identification

For participants who self-reported glaucoma during follow-up, we obtained their permission to retrieve their medical information. We asked diagnosing eye care providers to send all visual field (VF) reports and complete a glaucoma questionnaire about the participant’s maximum IOP, optic nerve features, gonioscopy findings and presence of XFM or other secondary causes for elevated IOP. In lieu of this request, eye care providers could send complete medical records. A glaucoma specialist (LRP) evaluated the information in a standardized manner to confirm diagnoses.

We defined XFG based on documentation of at least one of the following in the eye with XFM: 1) IOP >21 mm Hg, 2) cup-disc ratio ≥0.6 or 3) VF loss consistent with glaucoma on at least 1 recent reliable test. Participants who self-reported glaucoma and who were confirmed to have isolated elevated IOP or optic disc cupping, other forms of glaucoma (e.g., primary open-angle glaucoma, closed-angle glaucoma, other secondary glaucomas, etc.) or whose self-reports could not be confirmed (e.g., supporting medical records could not be obtained or the participant did not give permission to review records)37 were censored during follow-up as of the date of self-reported diagnosis.

Assessment of NMSC

From 1984 in NHS and from 1988 in HPFS, participants were asked about whether they had physician-diagnosed BCC or SCC, as two separate questions. For SCC, we included only self-reports that were confirmed after a review of medical and pathological reports (the SCC reports that could not be confirmed were censored as of the reported diagnosis date). For BCC, we included all self-reports because previous validation studies38–40 demonstrated a >90% accuracy of self-reported BCC, and the BCC validity was further demonstrated with an established genetic risk factor (melanocortin 1 receptor (MC1R) gene variant) being associated with BCC in this population.41

Prior to 1984 in NHS and 1988 in HPFS, participants were asked about whether they had either BCC or SCC as one single question (positive responses to this question are most likely to represent BCC history because 94% of the person-time attributed to NMSC after 1984 in NHS and 1988 in HPFS was BCC and 6% was SCC).

Thus, we defined having ‘any history of NMSC’ as any of 1) medical-record confirmed SCC (reported from 1984 in NHS and 1988 in HPFS), 2) self-reported BCC (from 1984 in NHS and 1988 in HPFS) or 3) self-reported combined SCC or BCC (prior to 1984 in NHS and 1988 in HPFS). If a participant had both SCC and BCC, then we assigned participants to the NMSC type that occurred first.

Assessment of predicted ambient erythemal UV radiation exposure

We evaluated ambient erythemal UV radiation exposure as a potential effect modifier. We hypothesized that the association between skin cancer and XFG may be stronger in areas with generally higher estimated ambient erythemal UV radiation exposure. Erythemal UV incorporates UV-A and UV-B wavelengths to calculate a measure describing the relative effectiveness of UV to induce erythema on Caucasian skin. We linked participants’ biennially updated geocoded residential addresses with an exposure model that predicted average July noon-time erythemal UV irradiance (mW/m2) at a 1 km2 spatial resolution and an annual temporal resolution from 1976 onwards in NHS and from 1986 onwards in HPFS.31 The UV exposure model31 was created by applying area-to-point residual kriging to downscale National Aeronautics and Space Administration (NASA) erythemal UV satellite remote sensing images and incorporated information on UV predictors (i.e., aerosol optical depth, cloud cover, dew point, elevation, latitude, ozone, surface albedo, surface incoming shortwave flux, and sulfur dioxide).31 Model cross-validation demonstrated high predictive performance compared to measurements at UV-B Monitoring and Research Program stations, showing positive percent relative improvements in mean absolute error (0.6–31.5%) and root mean square error (3.6–29.4%) in UV exposure prediction vs. using NASA satellite images only.

To capture participants’ long-term predicted ambient erythemal UV exposure, we calculated cumulatively-averaged values, where at each 2-year risk period, we averaged all of the available UV exposure information.

Statistical analysis

As the results from women and men did not show significant heterogeneity42 (e.g., p for heterogeneity≥0.65), we pooled the data from the two cohorts. We calculated the incidence by dividing incident cases of XFG by person-years for each exposure category. For multivariable-adjusted analyses, as a primary statistical approach, we conducted Cox proportional hazards analysis, while controlling for potential time-varying glaucoma risk factors to estimate multivariable-adjusted relative risks (MVRRs) and 95% confidence intervals (CIs). Analyses were performed in SAS 9.4 (SAS Institute, Cary NC).

In basic Cox regression models (i.e., Model 1), we evaluated just the exposure in relation to outcome and all analyses were stratified jointly by cohort, age in months, calendar year of each questionnaire cycle and latitude of residence. The stratifying factor of latitude of residence was defined as 5 categories of lifetime (birth to current) average continental US residence categorized by latitude (Northern only at > 42∘ N; Middle only at 37∘−42∘ N; Southern only at <37∘ N; residence between Northern and Middle States; and residence between Middle and Southern states). With such stratification, the univariate and multivariable-adjusted associations between NMSC and XFG are estimated within strata defined by age in months, calendar year and latitude. This approach finely and simultaneously adjusts for these variables as well as for any possible interactions of these variables. In multivariable-adjusted analyses, we stratified analyses by the same factors in Model 1 and included the following covariates (Model 2): glaucoma family history, major ancestry (Scandinavian Caucasian, Southern-European Caucasian, other ancestry), body mass index (kg/m2), cigarette smoking (pack-years), cumulatively updated intakes of alcohol (g/day), total calories (kcal/day), caffeine (mg/day),43 folate intake (μg/day),44 self-report (yes/no) of hypertension, diabetes, high cholesterol, and myocardial infarction, updated number of total reported eye exams and physician exams during follow-up. In Model 3, we further adjusted for quintiles of cumulatively-updated latitude17 based on biennially updated residential information. In Model 4, to have NMSC history best represent just personal UV exposure, we added the following covariates to Model 3 to further adjust for non-UV exposure related NMSC risk factors: vitamin A intake (IU/day),45 family history of melanoma (yes vs. no),46 natural hair color47 (red, blonde, light brown, dark brown, or black), number of arm moles (0,1–2, 3–5, or ≥6),48 sunburn susceptibility as a child/adolescent (no experience, no reaction/some redness, burn, or painful burn/blisters).49

Because the induction period between NMSC and XFG risk is not known, to better understand the temporal aspects of this relationship, we conducted analyses using NMSC history as of 4 years and 8 years prior to the 2-year risk period.50 To investigate effect modification for NMSC, we examined associations separately by family history of glaucoma, lifetime average continental US residence in northern states,17 ambient erythemal UV exposure, and age. We tested for effect modifications using Wald statistics of interaction terms in Cox regression models. All tests were 2-sided, with alpha of 0.05.

RESULTS

We identified 362 incident XFG cases in NHS and 83 cases in HPFS. At diagnosis, XFG cases were generally >65 years old (mean age at diagnosis was 68.3 years (SD=6.9) in women and 70.7 years (SD=7.1) in men), had high IOP (28.4 mmHg (SD=7.4)), and 54.8% of cases had only one eye with clinically evident XFS. At baseline, those who later became XFG cases were older by 4–5 years versus those who did not develop XFG (women in 1980: 50.0 years (SD=6.2) versus 45.9 (SD=6.9), respectively; men in 1986: 57.0 years (SD=9.5) versus 52.0 (SD=8.9), respectively).

Of the total person-time, NMSC history was prevalent in >10%; specifically, 9% were characterized with a history of BCC, 0.6% with SCC and 3.3% with early reports of combined NMSC. In both cohorts, any NMSC history (Table 1) was associated with being older, living in an area with higher ambient erythemal UV exposure, having a higher intake of alcohol and folate, having Scandinavian ancestry and high cholesterol, and having more eye exams and physician exams. However, NMSC history was associated with lower prevalence of diabetes, obesity, and living in the Northern tier. All these differences were accounted for in multivariable analyses.

Table 1.

Age and age-adjusted characteristics of total person-time accrued according to history of non-melanoma skin cancer in pooled data from two US cohorts: the Nurses’ Health Study (1980–2014) and the Health Professionals’ Follow-up Study (1986–2014)

WOMEN MEN
Non-melanoma skin cancer Non-melanoma skin cancer
No Yes No Yes
Total cohort-specific person-time (%) 89.2% 10.8% 81.4% 18.6%
Mean age (years) [SD] 58.8 [9.5] 66.8 [8.3] 60.6 [9.9] 67.8 [8.8]
Mean latitude (°N) [SD] 39.7 [3.6] 39.0 [4.0] 39.0 [4.9] 37.8 [5.3]
Residence in the northern tiera (%) 29.7 27.2 19.9 14.8
Mean erythemal UV exposure (mW/m2) [SD] 183.4 [24.9] 187.8 [27.7] 190.3 [28.8] 197.2 [31.1]
Scandinavian ancestry (%) 7.0 8.1 11.1 12.5
Family history of glaucoma (%) 14.1 14.9 12.3 12.9
Self-reported diabetes diagnosis (%) 6.1 5.7 6.3 5.5
Self-reported hypertension diagnosis (%) 37.4 37.5 35.2 35.6
Self-reported high cholesterol diagnosis (%) 41.2 45.2 42.4 44.9
Self-reported myocardial infarction (%) 1.9 1.7 6.6 6.1
≥30 pack-years of cigarette smoking (%) 16.8 16.2 15.3 14.1
Body mass index (kg/m2) ≥30 (%) 12.1 8.9 9.1 6.7
Mean total caloric intake (kcal/day) [SD] 1683.3 [435.9] 1713.7 [426.9] 1977.8 [549.8] 2011.0 [539.4]
Mean total folate (μg/day) [SD] 421.2 [200.3] 442.4 [195.3] 538.5 [248.5] 563.4 [249.4]
Mean alcohol intake (g/day) [SD] 5.9 [9.0] 6.9 [9.3] 10.8 [13.3] 11.8 [13.9]
Mean caffeine intake (g/day) [SD] 316.5 [212.6] 294.7 [199.0] 232.8 [216.5] 215.9 [203.4]
Mean number of eye exams reported (of 12) [SD] 6.4 [3.7] 7.0 [3.7] 5.5 [3.3] 5.9 [3.4]
Mean number of physician exams reported
(of 14) [SD]		 7.3 [4.1] 7.9 [4.0] 6.2 [3.8] 6.8 [3.9]
Open in a new tab
a

Lifetime residence in states north of the 41–42 degrees latitude north (AK, CT, ID, ME, MA, MI, MN, MT, NE, NH, NY, ND, OR, RI, SD, VT, WA, WI, WY)

In basic models (Table 2), NMSC was associated with 39% significantly increased risk of XFG (MVRR=1.39; 95% CI=1.08, 1.79). In multivariable-adjusted analyses, this was minimally attenuated; in Model 2, we observed a MVRR of 1.38 (95% CI=1.06, 1.78) for any NMSC history, and the estimate was identical in Model 3, where we addressed confounding by latitude. In Model 4 where we additionally adjusted for non-UV related NMSC risk factors, the MVRR did not change appreciably: MVRR of 1.40 (95% CI=1.08, 1.82) for any history of NMSC. Although we lacked the power to evaluate the subtypes of NMSC in relation to XFG, the MVRR estimates from Model 4 were 1.31 (95% CI=0.98, 1.75) for BCC; 2.41 (95% CI=0.98, 5.90) for SCC; and 1.60 (95% CI=0.95, 2.72) for early reports of combined NMSC.

Table 2.

Multivariable-adjusted relative risks (95% confidence intervals [CIs]) of non-melanoma skin cancer and the risk of exfoliation glaucoma in pooled data from two US cohorts: the Nurses’ Health Study (1980–2014) and the Health Professionals’ Follow-up Study (1986–2014)

No history Any history of non-melanoma skin cancera
Main analyses
Cases/Person-time (%) 347/1,683,297 (87.1%) 98/248,819 (12.9%)
Model 1: Basic modelb 1.00 (reference) 1.39 (1.08, 1.79)
Model 2: Core covariatesc 1.00 (reference) 1.38 (1.06, 1.78)
Model 3: Model 2 + latituded 1.00 (reference) 1.38 (1.06, 1.78)
Model 4: Model 3 + non-melanoma skin cancer risk factors unrelated to UV exposuree 1.00 (reference) 1.40 (1.08, 1.82)
4-year lag
Cases/Person-time (%) 374/1,747,279 (90.4%) 71 / 184,837 (9.6%)
Model 1: Basic modelb 1.00 (reference) 1.35 (1.01, 1.79)
Model 2: Core covariatesc 1.00 (reference) 1.36 (1.02, 1.82)
Model 3: Model 2 + latituded 1.00 (reference) 1.36 (1.02, 1.82)
Model 4: Model 3 + non-melanoma skin cancer risk factors unrelated to UV exposuree 1.00 (reference) 1.38 (1.03, 1.84)
8-year lag
Cases/Person-time (%) 391/1,799,766 (93.1%) 54 / 132,350 (9.6%)
Model 1: Basic modelb 1.00 (reference) 1.39 (1.01, 1.91)
Model 2: Core covariatesc 1.00 (reference) 1.43 (1.03, 1.97)
Model 3: Model 2 + latituded 1.00 (reference) 1.43 (1.03, 1.97)
Model 4: Model 3 + non-melanoma skin cancer risk factors unrelated to UV exposuree 1.00 (reference) 1.45 (1.04, 2.01)
No history Non-melanoma skin cancer at age <55 years Non-melanoma skin cancer at age ≥55 years
Non-melanoma skin cancer age at diagnosis
Cases/Person-time (%) 347/1,683,297 (87.1%) 31/108,680 (5.6%) 67/140,139 (7.3%)
Model 1: Basic modelb 1.00 (reference) 1.60 (1.06, 2.40) 1.30 (0.97, 1.75)
Model 2: Core covariatesc 1.00 (reference) 1.59 (1.05, 2.41) 1.29 (0.95, 1.74)
Model 3: Model 2 + latituded 1.00 (reference) 1.59 (1.05, 2.41) 1.29 (0.95, 1.74)
Model 4: Model 3 + non-melanoma skin cancer risk factors unrelated to UV exposuree 1.00 (reference) 1.62 (1.07, 2.46) 1.32 (0.97, 1.78)
Open in a new tab
a

Defined as any self-report of basal cell carcinoma (from 1984 in women and from 1988 in men), any confirmed squamous cell carcinoma (from 1984 in women and from 1988 in men), or any self-report of any non-melanoma skin cancer (before 1984 in women and before 1988 in men)

b

For model 1, analyses were stratified by cohort, age in months, 2-year period at risk, and 5 categories of lifetime (birth to current) residential history geographical tier: lifetime residence in states north of the 41–42 degrees latitude north (AK, CT, ID, ME, MA, MI, MN, MT, NE, NH, NY, ND, OR, RI, SD, VT, WA, WI, WY); lifetime residence in the states south of the 37 degrees latitude north (AL, AZ, AR, FL, GA, HI, LA, MS, NM, NC, OK, PR, SC, TN, TX and southern California from Los Angeles to its southern border); lifetime residence in 37–40 degrees latitude north (CO, DE, DC, IL, IN, IA, KS, KY, MD, MO, NV, NJ, OH, PA, UT, VA, WV), including the remainder of California north of Los Angeles; lifetime residence between northern and middle states; lifetime residence between middle and southern states.

c

For model 2, analyses were stratified by the same variables as in model 1 and in addition, analyses were adjusted for ancestry (Scandinavian Caucasian, Southern-European Caucasian, other), glaucoma family history, high cholesterol, hypertension, diabetes, myocardial infarction, body mass index (22–23, 24–25, 26–27, 28–29, 30+ kg/m2), pack-years of cigarette smoking (1–9, 10–19, 20–29, 30+ pack-years), cumulatively averaged total energy intake (kcal/day), alcohol intake (g/day), caffeine intake (mg/day), folate intake (μg/day), cumulative number of eye exams reported (linear count), cumulative number of physician exams reported (linear count)

d

Model 3 was model 2 with additional adjustment for cumulatively updated latitude from baseline (as a continuous variable)

e

Model 4 was model 3 with additional adjustment for non-UV exposure related risk factors for non-melanoma skin cancer including vitamin A intake (IU/day), family history of melanoma (yes vs. no), natural hair color (red, blonde, light brown, dark brown, or black), number of arm moles (0,1–2, 3–5, or ≥6), sunburn susceptibility as a child/adolescent (no experience, no reaction/some redness, burn, or painful burn/blisters)

To evaluate the temporal aspects of the relationship between NMSC and XFG, we evaluated NMSC history as of ~ 4 or ~ 8 years prior to diagnosis versus the most updated history. The 4-year lagged (Table 2; MVRR from Model 4=1.38; 95% CI=1.03, 1.84) and 8-year lagged analyses (MVRR from Model 4=1.45; 95%CI=1.04, 2.01) showed similar associations as the main analyses, indicating a relatively long induction period. We evaluated NMSC history stratified by age at diagnosis (<55 years old and age ≥55 years; Table 2): the MVRR (from Model 4) for history of NMSC that occurred before age 55 in relation to XFG was 1.62 (95% CI=1.07, 2.46), while the MVRR for a history of NMSC on or after age 55 years was 1.32 (95% CI=0.97, 1.78). To indirectly test the difference between these two estimates, we restricted analyses to only those with NMSC history and observed that the difference in time to outcome between the two age groups was not significant (p=0.27).

While no significant interaction was observed with glaucoma family history (Table 3; p for interaction=0.84), we observed that the association between NMSC and XFG was significantly stronger in those younger than 65 years versus those 65 years or older (p for interaction=0.01; MVRR from Model 4=2.56; 95% CI=1.62, 4.05 among those aged <65 years versus MVRR=1.25; 95% CI=0.94, 1.66 among those aged ≥65 years). Also, compared to those living in lower tiers, the association was significantly stronger in those whose lifetime residential history was in the northern tier (defined as north of the 41–42 degrees latitude north or in the US states of AK, CT, ID, ME, MA, MI, MN, MT, NE, NH, NY, ND, OR, RI, SD, VT, WA, WI, WY; p for interaction=0.04; MVRR=1.92; 95% CI=1.28, 2.88 among those living in the northern tier versus MVRR=1.19; 95% CI=0.86, 1.66 among those living in lower tiers). Finally, a non-significant difference was observed by ambient erythemal UV exposure (p for interaction=0.09), where NMSC history was more strongly associated with XFG among those living in areas with higher ambient erythemal UV exposure within the same residential tier (MVRR=1.93; 95% CI=1.33, 2.81) versus those living in areas with lower exposure (MVRR=0.98; 95% CI=0.65, 1.48).

Table 3.

Multivariable-adjusted relative risks (95% confidence intervals)a for history of non-melanoma skin cancer by residence history, family history of glaucoma, and age in relation to the risk of exfoliation glaucoma

No history of non-melanoma skin cancer Any history of non-melanoma skin cancer combineda P for interaction
By lifetime residence
 <42° north (middle or southern tier; 73.3%)b 1.00 (reference) 1.19 (0.86, 1.66)
 ≥42° north (northern tier; 26.7%)b 1.00 (reference) 1.92 (1.28, 2.88) 0.04
By ambient erythemal UV exposure
 < median of 175 mW/m2; (50.0%) 1.00 (reference) 0.98 (0.65, 1.48)
 ≥ median of 175 mW/m2; (50.0%) 1.00 (reference) 1.93 (1.33, 2.81) 0.09
By family history of glaucoma
 Without family history (86.3%) 1.00 (reference) 1.37 (1.00, 1.87)
 With family history (13.7%) 1.00 (reference) 2.08 (0.87, 4.99) 0.84
By age
 <65 years (35.0%) 1.00 (reference) 2.56 (1.62, 4.05)
 ≥65 years (65.0%) 1.00 (reference) 1.25 (0.94, 1.66) 0.01
Open in a new tab
a

Multivariable analyses were stratified for and adjusted for the same covariates as listed for Model 4 in footnotee of Table 2.

b

States that are ≥42° latitude north are provided in footnoteb of Table 2.

DISCUSSION

In this large prospective study of 120,307 participants followed for 28+ years, any NMSC history, either BCC or SCC, was associated with higher XFG risk. Because BCC and SCC result, in part, from increased dermal exposure to UV radiation and are typically distributed in the head neck region,51 they can be considered markers of higher personal ocular UV radiation exposure. The adverse association observed with NMSC history indirectly supports our hypothesis that greater personal ocular UV radiation exposure may increase XFG risk, and the association observed with remote NMSC history (shown in 8-year lagged analysis) indicates a long latency in the relation with XFG.

Measuring lifetime ocular UV radiation exposure is challenging as the pattern of ocular exposure may be different from skin exposure and dependent on a myriad of factors. For example, depending on the season and latitude, ocular UV exposure can peak in the early morning and afternoon versus at noon for skin exposure.19,52 Also, ocular exposure is also affected by the cornea absorbing most of the UVB rays,53 orbital anatomies, squinting, pupil constriction and use of eye protection such as sunglasses and wide-brimmed hats.54 Thus, we evaluated history of NMSC diagnosis as an indirect measure of ocular UV radiation, as NMSC history has been strongly associated with measures of UV flux previously in our study populations.26,55,56

Consistent with a study that in univariate analyses observed that skin cancer of any type was associated with XFS,18 we observed that the risk of XFG was 1.40 fold higher with NMSC. Because we controlled for various non-UV-related risk factors for NMSC in multivariable-adjusted analyses, variation in NMSC history would be strongly correlated with personal cumulative ocular UV exposure. This is also supported by the fact that NMSC lesions are typically distributed in the head and neck region,51 and 5%–10% of all NMSCs occur on the eyelids.57

Also, BCC and SCC history may represent greater childhood UV exposure, as greater blistering sunburns in childhood and adolescence was associated with BCC and SCC in the younger NHS II cohort;26 and this observation would be in line with a previous report of greater time spent outdoors during the teenage/young adult years being associated with a doubling of XFG risk.58 This, combined with the fact that even 8-year lagged analyses showed adverse associations imply that there is a prolonged latency phase.

Furthermore, we observed stronger associations between NMSC and XFG among those whose lifetime residence was in the Northern US states. UV exposure to the eye is likely greater in non-equatorial latitudes,17–19,23 especially close to the polar regions where ultraviolet intensity is high.20,21 This is likely because the sun’s rays come into the eye at a lower angle in the sky throughout the day (~ 40◦ angle)59 versus direct vertical sunlight exposure on the skin (which would be highest in equatorial regions). Also, other sources of UV exposure such as scattered light and reflected light may arguably be more important for the eye, and non-equatorial regions allow for more reflected ocular UV exposure as they are more likely to have snow, which can reflect up to 80% of the sun’s rays.22 Thus, those with NMSC history may benefit from regular eye exams to detect XFS/XFG earlier—particularly for younger participants aged 40–64 years living in northern latitudes in whom we observed significantly stronger associations with NMSC. Clearly, further research is needed to better understand the clinical correlations between NMSC and XFS/XFG as well as any common etiologic pathways (e.g., lysyl oxidases being upregulated with UV exposure in in vitro studies models using both human eye60 and skin tissues61,62).

This work has several limitations. First, because the study population was large, participants lived across the US, we were not able to conduct direct standardized regular eye exams; hence, we had low sensitivity to detect cases of XFS/XFG, and we are unable to provide estimates of absolute XFG risks or estimates of differences in absolute risk. However, as proven in epidemiological methods literature,63 we were able to validly estimate relative risks in this setting, given that we had a very specific definition of the outcome and ascertainment of disease was not related to exposure (we adjusted for both cumulative number of reported eye exams and physician exams during the 25+ years of follow-up). Second, we were not able to evaluate cutaneous melanoma and XFG, but melanoma is more strongly associated with discontinuous solar exposure,64 while BCC and SCC are keratinocyte cancers that are linked to regular ongoing solar exposure,55 a pattern of exposure that may be more associated with XFG. Third, there is always a possibility of detection bias where those who are diagnosed with BCC and SCC may be screened more frequently for health in general and thus are more likely to be diagnosed with XFG; this possibility was likely diminished given that we restricted follow-up to only those reporting eye exams, and we further adjusted for the updated number of self-reported eye exams and physician exams during follow-up (both metrics of greater health screenings). Finally, because our study participants were predominantly Caucasian, our results may not be applicable to minority populations with different underlying XFG risk. This is especially applicable to people of Asian and African descent, where BCC and SCC are less common with more atypical presentations and cutaneous distributions.65

In conclusion, a history of NMSC, a marker of greater personal UV radiation exposure, was associated with higher XFG risk. If confirmed, our results would provide further support for greater eye protection on sunny days66,67 and regular eye exams for those aged 40+ years, particularly those with NMSC history.

ACKNOWLEDGMENTS

The authors assume full responsibility for analyses and interpretation of these data.

We would like to thank the participants and staff of the NHS and HPFS for their valuable contributions as well as the following state cancer registries for their help: AL, AZ, AR, CA, CO, CT, DE, FL, GA, ID, IL, IN, IA, KY, LA, ME, MD, MA, MI, NE, NH, NJ, NY, NC, ND, OH, OK, OR, PA, RI, SC, TN, TX, VA, WA, WY.

Financial Support: This work was supported by grants UM1 CA186107, U01 CA167552, P01 CA87969, T32 CA009001, R01 EY020928, R01 EY015473 and R01 EY09611 from the National Institutes of Health. Dr. Louis Pasquale is also supported by the Eye and Vision Research Institute of the Icahn School of Medicine at Mount Sinai. The sponsor(s) or funding organization(s) had no role in the design or conduct of this research.

Footnotes

Conflict of Interest: Unrelated to this work, Dr. Qureshi has received honoraria from AbbVie, Amgen, the Centers for Disease Control and Prevention, Janssen, Merck, Pfizer, and Novartis (as a consultant), and he is an investigator (without financial compensation) for Sanofi and Regeneron. Unrelated to this work, Dr. Pasquale is a consultant for Bausch & Lomb, Verily, Nicox, Emerald Bioscience and Eyenovia. Unrelated to this work, Dr. Wiggs is a consultant for Aerpio, Maze, Allergan, Editas and Regenxbio. There are no financial disclosures for the other authors.

REFERENCES

  • 1.Ocular Ritch R. and systemic manifestations of exfoliation syndrome. J Glaucoma. 2014;23:S1–8. [DOI] [PubMed] [Google Scholar]
  • 2.Ritch R Exfoliation syndrome and occludable angles. Trans Am Ophthalmol Soc. 1994;92:845–944. [PMC free article] [PubMed] [Google Scholar]
  • 3.Kanthan GL, Mitchell P, Burlutsky G, et al. Pseudoexfoliation syndrome and the long-term incidence of cataract and cataract surgery: the blue mountains eye study. Am J Ophthalmol. 2013;155:83–88 e81. [DOI] [PubMed] [Google Scholar]
  • 4.Davis D, Brubaker J, Espandar L, et al. Late in-the-bag spontaneous intraocular lens dislocation: evaluation of 86 consecutive cases. Ophthalmology. 2009;116:664–670. [DOI] [PubMed] [Google Scholar]
  • 5.Karagiannis D, Kontadakis GA, Klados NE, et al. Central retinal vein occlusion and pseudoexfoliation syndrome. Clin Interv Aging. 2015;10:879–883. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Cursiefen C, Hammer T, Kuchle M, et al. Pseudoexfoliation syndrome in eyes with ischemic central retinal vein occlusion. A histopathologic and electron microscopic study. Acta Ophthalmol. 2001;79:476–478. [DOI] [PubMed] [Google Scholar]
  • 7.Ritch R, Prata TS, de Moraes CG, et al. Association of exfoliation syndrome and central retinal vein occlusion: an ultrastructural analysis. Acta Ophthalmol. 2010;88:91–95. [DOI] [PubMed] [Google Scholar]
  • 8.Rasmussen CA, Kaufman PL. The trabecular meshwork in normal eyes and in exfoliation glaucoma. J Glaucoma. 2014;23:S15–19. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Thorleifsson G, Magnusson KP, Sulem P, et al. Common sequence variants in the LOXL1 gene confer susceptibility to exfoliation glaucoma. Science. 2007;317:1397–1400. [DOI] [PubMed] [Google Scholar]
  • 10.Poli MC, Ebstein F, Nicholas SK, et al. Heterozygous Truncating Variants in POMP Escape Nonsense-Mediated Decay and Cause a Unique Immune Dysregulatory Syndrome. Am J Hum Genet. 2018;102:1126–1142. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Aung T, Ozaki M, Lee MC, et al. Genetic association study of exfoliation syndrome identifies a protective rare variant at LOXL1 and five new susceptibility loci. Nat Genet. 2017;49:993–1004. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Forsman E, Cantor RM, Lu A, et al. Exfoliation syndrome: prevalence and inheritance in a subisolate of the Finnish population. Acta Ophthalmol. 2007;85:500–507. [DOI] [PubMed] [Google Scholar]
  • 13.Topouzis F, Wilson MR, Harris A, et al. Prevalence of open-angle glaucoma in Greece: the Thessaloniki Eye Study. Am J Ophthalmol. 2007;144:511–519. [DOI] [PubMed] [Google Scholar]
  • 14.Rao RQ, Arain TM, Ahad MA. The prevalence of pseudoexfoliation syndrome in Pakistan. Hospital based study. BMC Ophthalmol. 2006;6:27. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Colin J, Le Gall G, Le Jeune B, et al. The prevalence of exfoliation syndrome in different areas of France. Acta Ophthalmol Suppl. 1988;184:86–89. [DOI] [PubMed] [Google Scholar]
  • 16.Rudkin AK, Edussuriya K, Sennanayake S, et al. Prevalence of exfoliation syndrome in central Sri Lanka: the Kandy Eye Study. Br J Ophthalmol. 2008;92:1595–1598. [DOI] [PubMed] [Google Scholar]
  • 17.Kang JH, Loomis S, Wiggs JL, et al. Demographic and geographic features of exfoliation glaucoma in 2 United States-based prospective cohorts. Ophthalmology. 2012;119:27–35. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Stein JD, Pasquale LR, Talwar N, et al. Geographic and climatic factors associated with exfoliation syndrome. Arch Ophthalmol. 2011;129:1053–1060. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Sasaki H, Sakamoto Y, Schnider C, et al. UV-B exposure to the eye depending on solar altitude. Eye Contact Lens. 2011;37:191–195. [DOI] [PubMed] [Google Scholar]
  • 20.McKenzie R, Connor B, Bodeker G. Increased summertime UV radiation in New Zealand in response to ozone loss. Science. 1999;285:1709–1711. [DOI] [PubMed] [Google Scholar]
  • 21.Petkov BH, Vitale V, Tomasi C, et al. Response of the ozone column over Europe to the 2011 Arctic ozone depletion event according to ground-based observations and assessment of the consequent variations in surface UV irradiance. Atmos Environ. 2014;85:169–178. [Google Scholar]
  • 22.Sliney DH. Physical factors in cataractogenesis: ambient ultraviolet radiation and temperature. Invest Ophthalmol Vis Sci. 1986;27:781–790. [PubMed] [Google Scholar]
  • 23.Pasquale LR, Jiwani AZ, Zehavi-Dorin T, et al. Solar exposure and residential geographic history in relation to exfoliation syndrome in the United States and Israel. JAMA Ophthalmol. 2014;132:1439–1445. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Marshall J. Radiationand the Ageing Eye. Ophthal Physiol Opt. 1984;4:1–23. [Google Scholar]
  • 25.Lai V, Cranwell W, Sinclair R. Epidemiology of skin cancer in the mature patient. Clin Dermatol. 2018;36:167–176. [DOI] [PubMed] [Google Scholar]
  • 26.Wu S, Han J, Laden F, et al. Long-term ultraviolet flux, other potential risk factors, and skin cancer risk: a cohort study. Cancer Epidemiol Biomarkers Prev. 2014;23:1080–1089. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Scotto J, Fears TR, Fraumeni JF Jr. Solar radiation In: Schottenfeld DFJ Jr., ed. Cancer Epidemiology and Prevention. New York: Oxford University Press; 1996:355–372. [Google Scholar]
  • 28.Scotto J, Fears TR. The association of solar ultraviolet and skin melanoma incidence among caucasians in the United States. Cancer Invest. 1987;5:275–283. [PubMed] [Google Scholar]
  • 29.Fears TR, Bird CC, Guerry Dt, et al. Average midrange ultraviolet radiation flux and time outdoors predict melanoma risk. Cancer Res. 2002;62:3992–3996. [PubMed] [Google Scholar]
  • 30.NTP (National Toxicology Program). Report on Carcinogens, Fourteenth Edition In. Research Triangle Park: U.S. Department of Health and Human Services; 2016. [Google Scholar]
  • 31.VoPham T, Hart JE, Bertrand KA, et al. Spatiotemporal exposure modeling of ambient erythemal ultraviolet radiation. Environ Health. 2016;15:111. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Li WQ, Cho E, Wu S, et al. Host Characteristics and Risk of Incident Melanoma by Breslow Thickness. Cancer Epidemiol Biomarkers Prev. 2019;28:217–224. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Liang G, Nan H, Qureshi AA, et al. Pre-diagnostic plasma 25-hydroxyvitamin D levels and risk of non-melanoma skin cancer in women. PLoS One. 2012;7:e35211. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Garcia VC, Martini LA. Vitamin D and cardiovascular disease. Nutrients. 2010;2:426–437. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Barton J, Bain C, Hennekens CH, et al. Characteristics of respondents and non-respondents to a mailed questionnaire. Am J Public Health. 1980;70:823–825. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Grobbee DE, Rimm EB, Giovannucci E, et al. Coffee, caffeine, and cardiovascular disease in men. N Engl J Med. 1990;323:1026–1032. [DOI] [PubMed] [Google Scholar]
  • 37.Kang JH, Ivey KL, Boumenna T, et al. Prospective study of flavonoid intake and risk of primary open-angle glaucoma. Acta Ophthalmol. 2018;96:e692–e700. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.van Dam RM, Huang Z, Giovannucci E, et al. Diet and basal cell carcinoma of the skin in a prospective cohort of men. Am J Clin Nutr. 2000;71:135–141. [DOI] [PubMed] [Google Scholar]
  • 39.Colditz GA, Martin P, Stampfer MJ, et al. Validation of questionnaire information on risk factors and disease outcomes in a prospective cohort study of women. Am J Epidemiol. 1986;123:894–900. [DOI] [PubMed] [Google Scholar]
  • 40.Hunter DJ, Colditz GA, Stampfer MJ, et al. Diet and risk of basal cell carcinoma of the skin in a prospective cohort of women. Ann Epidemiol. 1992;2:231–239. [DOI] [PubMed] [Google Scholar]
  • 41.Han J, Kraft P, Colditz GA, et al. Melanocortin 1 receptor variants and skin cancer risk. Int J Cancer. 2006;119:1976–1984. [DOI] [PubMed] [Google Scholar]
  • 42.DerSimonian R, Laird N. Meta-Analysis in Clinical Trials. Control Clin Trials. 1986;7:177–188. [DOI] [PubMed] [Google Scholar]
  • 43.Pasquale LR, Wiggs JL, Willett WC, et al. The Relationship between caffeine and coffee consumption and exfoliation glaucoma or glaucoma suspect: a prospective study in two cohorts. Invest Ophthalmol Vis Sci. 2012;53:6427–6433. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44.Kang JH, Loomis SJ, Wiggs JL, et al. A prospective study of folate, vitamin B(6), and vitamin B(1)(2) intake in relation to exfoliation glaucoma or suspected exfoliation glaucoma. JAMA Ophthalmol. 2014;132:549–559. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 45.Samarawickrama C, Chew S, Watson S. Retinoic acid and the ocular surface. Surv Ophthalmol. 2015;60:183–195. [DOI] [PubMed] [Google Scholar]
  • 46.Berlin NL, Cartmel B, Leffell DJ, et al. Family history of skin cancer is associated with early-onset basal cell carcinoma independent of MC1R genotype. Cancer Epidemiol. 2015;39:1078–1083. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 47.Grodstein F, Speizer FE, Hunter DJ. A prospective study of incident squamous cell carcinoma of the skin in the nurses’ health study. J Natl Cancer Inst. 1995;87:1061–1066. [DOI] [PubMed] [Google Scholar]
  • 48.Wu S, Han J, Li WQ, et al. Basal-cell carcinoma incidence and associated risk factors in U.S. women and men. Am J Epidemiol. 2013;178:890–897. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 49.Wu S, Cho E, Li WQ, et al. History of Severe Sunburn and Risk of Skin Cancer Among Women and Men in 2 Prospective Cohort Studies. Am J Epidemiol. 2016;183:824–833. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 50.Willett WC. Nutritional Epidemiology. Third ed. New York: Oxford University Press, Inc.; 2013. [Google Scholar]
  • 51.Subramaniam P, Olsen CM, Thompson BS, et al. Anatomical Distributions of Basal Cell Carcinoma and Squamous Cell Carcinoma in a Population-Based Study in Queensland, Australia. JAMA Dermatol. 2017;153:175–182. [DOI] [PubMed] [Google Scholar]
  • 52.Sydenham MM, Collins MJ, Hirst LW. Measurement of ultraviolet radiation at the surface of the eye. Invest Ophthalmol Vis Sci. 1997;38:1485–1492. [PubMed] [Google Scholar]
  • 53.Calhoun WR, Akpek EK, Weiblinger R, et al. Evaluation of broadband spectral transmission characteristics of fresh and gamma-irradiated corneal tissues. Cornea. 2015;34:228–234. [DOI] [PubMed] [Google Scholar]
  • 54.Behar-Cohen F, Baillet G, de Ayguavives T, et al. Ultraviolet damage to the eye revisited: eye-sun protection factor (E-SPF(R)), a new ultraviolet protection label for eyewear. Clin Ophthalmol. 2014;8:87–104. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 55.Wu S, Han J, Vleugels RA, et al. Cumulative ultraviolet radiation flux in adulthood and risk of incident skin cancers in women. Br J Cancer. 2014;110:1855–1861. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 56.Qureshi AA, Laden F, Colditz GA, et al. Geographic variation and risk of skin cancer in US women. Differences between melanoma, squamous cell carcinoma, and basal cell carcinoma. Arch Intern Med. 2008;168:501–507. [DOI] [PubMed] [Google Scholar]
  • 57.Cook BE, Jr., Bartley GB. Epidemiologic characteristics and clinical course of patients with malignant eyelid tumors in an incidence cohort in Olmsted County, Minnesota. Ophthalmology. 1999;106:746–750. [DOI] [PubMed] [Google Scholar]
  • 58.Kang JH, Wiggs JL, Pasquale LR. Relation between time spent outdoors and exfoliation glaucoma or exfoliation glaucoma suspect. Am J Ophthalmol. 2014;158:605–614 e601. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 59.Hu L, Gao Q, Gao N, et al. Solar UV exposure at eye is different from environmental UV: diurnal monitoring at different rotation angles using a manikin. J Occup Environ Hyg. 2013;10:17–25. [DOI] [PubMed] [Google Scholar]
  • 60.Zenkel M, Schlotzer-Schrehardt U. Expression and regulation of LOXL1 and elastin-related genes in eyes with exfoliation syndrome. J Glaucoma. 2014;23:S48–50. [DOI] [PubMed] [Google Scholar]
  • 61.Langton AK, Tsoureli-Nikita E, Griffiths CEM, et al. Lysyl oxidase activity in human skin is increased by chronic ultraviolet radiation exposure and smoking. Br J Dermatol. 2017;176:1376–1378. [DOI] [PubMed] [Google Scholar]
  • 62.Majora M, Wittkampf T, Schuermann B, et al. Functional consequences of mitochondrial DNA deletions in human skin fibroblasts: increased contractile strength in collagen lattices is due to oxidative stress-induced lysyl oxidase activity. Am J Pathol. 2009;175:1019–1029. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 63.Rothman KJ, Greenland S. Modern Epidemiology. Second ed. Philadelphia, PA: Lippincott-Raven Publishers; 1998. [Google Scholar]
  • 64.Gandini S, Sera F, Cattaruzza MS, et al. Meta-analysis of risk factors for cutaneous melanoma: II. Sun exposure. Eur J Cancer. 2005;41:45–60. [DOI] [PubMed] [Google Scholar]
  • 65.Higgins S, Nazemi A, Chow M, et al. Review of Nonmelanoma Skin Cancer in African Americans, Hispanics, and Asians. Dermatol Surg. 2018;44:903–910. [DOI] [PubMed] [Google Scholar]
  • 66.Protecting children from ultraviolet radiation - archived, 11 December 2009. https://www.who.int/uv/resources/archives/fs261/en/. Accessed 3/10/2019.
  • 67.Prevent Blindness America. Children’s Eyes More Susceptible to Long-Term Damage from UV Rays. https://www.preventblindness.org/childrens-eyes-are-more-susceptible-long-term-damage-uv-rays-0. Accessed 07/15/2019.

RESOURCES