Pasquale LR, Jiwani AZ, Zehavi-Dorin T, Majd A, Rhee DJ, Chen T, Turalba A, Shen L, Brauner S, Grosskreutz C, Gardiner M, Chen S, Borboli-Gerogiannis S, Greenstein SH, Chang K, Ritch R, Loomis S, Kang JH, Wiggs JL, Levkovitch-Verbin H. Solar Exposure and Residential Geographic History in Relation to Exfoliation Syndrome in the United States and Israel. JAMA Ophthalmol. 2014;132(12):1439-1445. doi:10.1001/jamaophthalmol.2014.3326
Residential (geographic) history and extent of solar exposure may be important risk factors for exfoliation syndrome (XFS) but, to our knowledge, detailed lifetime solar exposure has not been previously evaluated in XFS.
To assess the relation between residential history, solar exposure, and XFS.
Design, Setting, and Participants
This clinic-based case-control study was conducted in the United States and Israel. It involved XFS cases and control individuals (all ≥60-year-old white individuals) enrolled from 2010 to 2012 (United States: 118 cases and 106 control participants; Israel: 67 cases and 72 control participants).
Main Outcomes and Measures
Weighted lifetime average latitude of residence and average number of hours per week spent outdoors as determined by validated questionnaires.
In multivariable analyses, each degree of weighted lifetime average residential latitude away from the equator was associated with 11% increased odds of XFS (pooled odds ratio [OR], 1.11; 95% CI, 1.05-1.17; P < .001). Furthermore, every hour per week spent outdoors during the summer, averaged over a lifetime, was associated with 4% increased odds of XFS (pooled OR, 1.04; 95% CI, 1.00-1.07; P = .03). For every 1% of average lifetime summer time between 10 am and 4 pm that sunglasses were worn, the odds of XFS decreased by 2% (OR, 0.98; 95% CI, 0.97-0.99; P < .001) in the United States but not in Israel (OR, 1.00; 95% CI, 0.99-1.01; P = .92; P for heterogeneity = .005). In the United States, after controlling for important environmental covariates, history of work over water or snow was associated with increased odds of XFS (OR, 3.86; 95% CI, 1.36-10.9); in Israel, there were too few people with such history for analysis. We did not identify an association between brimmed hat wear and XFS (P > .57).
Conclusions and Relevance
Lifetime outdoor activities may contribute to XFS. The association with work over snow or water and the lack of association with brimmed hat wear suggests that ocular exposure to light from reflective surfaces may be an important type of exposure in XFS etiology.
Exfoliation syndrome (XFS) is a form of deleterious ocular aging categorized by increased risk for climatic droplet keratopathy,1,2 age-related cataract,3- 5 late-onset spontaneous intraocular lens subluxation,6 elevated intraocular pressure with glaucomatous optic neuropathy,7 and retinal vein occlusion.8,9 The discovery that common variants of LOXL1 are present in 99% of XFS cases represents a significant advance in understanding this condition.10 However, 80% of control individuals also harbor these variants and the ratio of cases to control individuals with trait-related variants is fairly similar in regions where XFS is hyperendemic11 and regions where the condition is relatively rare.12 This suggests that other genetic or environmental factors contribute to XFS.
There is considerable evidence that climatic factors contribute to XFS. For example, aboriginal Australians who spend substantial time outdoors have a higher prevalence of XFS (11%)13 compared with a mix of urban, rural, and nursing home residents from Victoria, Australia (1%).14 Taylor15 found that XFS was particularly common in Australian stockmen who looked after livestock. On the island of Rab in the Adriatic Sea, the frequency of XFS was 23% among 480 villagers categorized as agriculturists and fishermen but no cases were detected in 61 urban dwellers.16 Stein et al17 concluded that people with XFS tended to reside at higher latitudes in the continental United States, a finding that was supported by the Nurses’ Health Study and Health Professionals Follow-up Study.18 When state-specific climatic data were considered, colder ambient temperature and an increased number of sunny days was associated with increased risk for XFS.17
The relation between ultraviolet (UV) radiation (UVR) and XFS deserves further study because some reports have not been consistent with a positive association between UVR and XFS. For instance, Forsius et al19 found virtually no XFS among Inuit people residing in Greenland, where UVR exposure is high. Forsius et al19 did find at least 1 type of solar ophthalmopathy—specifically, pterygium, climatic keratopathy, or pronounced pingecula—that was at least as common, or more common, in individuals with XFS compared with unaffected individuals in various populations. Yet, because the frequency of XFS in tropical countries (Tunisia and India) lagged behind that in countries at higher latitudes (Finland, Iceland, and Russia) in their worldwide survey, they concluded that “climate does not appear to influence the occurrence of XFS.”19 Finally, no relation between time spent outdoors and the risk for XFS was found in the Reykjavik Eye Study.20
To further clarify the UVR-XFS relationship, we conducted a clinic-based, case-control study in the United States and Israel. We administered validated questionnaires to explore the relation between residential (geographic) history and solar exposure from birth to age 60 years in relation to XFS.
We conducted a clinic-based, case-control study at 2 ophthalmic centers: Massachusetts Eye and Ear Infirmary in the United States and the Goldschleger Eye Institute in Israel. The Israeli site was chosen because we wanted to assess the previously reported association between increasing latitude and XFS in Europe and the Goldschleger Eye Institute cares for people from throughout Europe. The human subject committee at each institution approved this study, and written informed consent was obtained from participants.
All participants were at least 60 years old and recruited from November 2, 2010, to December 20, 2012. Cases had evidence of exfoliation precipitates in at least 1 eye with or without evidence of glaucoma. Control participants had no exfoliation material on any available examinations including at least 1 dilated slitlamp examination prior to cataract surgery in both eyes if the participant was pseudophakic. Control participants could have forms of glaucoma other than exfoliation glaucoma. We collected data regarding a family history of glaucoma, diabetes mellitus, hypertension, and eye color from medical record review.
Measurement of residential history and solar exposure were based on the Residential History and Ocular Exposure to UV Light instruments found in the PhenX Toolkit,21 a repository of freely available validated instruments designed to assess traits related to complex disease (http://www.phenxtoolkit.org). The questionnaires were administered by telephone or in person by trained interviewers (A.Z.J., T.Z.-D., and A.M.), who were masked to participants’ ophthalmic status.
For residential history, we asked the participants where they lived from birth to age 60 years, capturing all moves to new locations. We recorded the latitude of each residence with a Google map application prior to its retirement in August 2013 (http://maps.google.com/latitude). We calculated the weighted lifetime average latitude from birth to age 60 years, which accounted for the time spent at each residential location.
We asked participants to provide information on the activities listed here during the age periods of 10 to 19, 20 to 29, 30 to 39, 40 to 49, and 50 to 59 years:
The number of hours per week spent outdoors between 10 am and 4 pm during the summer and the percentage of that time eyeglasses, sunglasses (including clip on, wrap around, or other forms of ocular UV protection), and brimmed hats or visors were worn.
Whether time was spent in regular leisure activity over snow or a body of water (such as the ocean, lake, or pond). We defined regular leisure activity as an activity performed at least 1 whole week in the aggregate (or the equivalent of engaging in an activity from 10 am to 4 pm for the equivalent of 7 days or approximately 42 hours) during a year.
Whether time was spent working over snow in activities such as ski instructing or winter landscaping (snow shoveling for personal housekeeping was excluded because it does not constitute a regular environmental exposure).
As an ancillary component to the solar exposure questionnaire, we asked about the age when participants experienced their first sunburn, defined as a solar exposure that caused redness and pain for approximately 12 hours or more, even if it only involved a small patch of skin.
We used SAS version 9.3 (SAS Institute) for all statistical analyses. We performed multiple logistic regression models separately for the US and Israeli groups. In our various models, we adjusted for sex, age in years, light eye color (ie, hazel, green, blue green, blue, or gray), diabetes mellitus, hypertension, family history of glaucoma, lifetime average number of hours spent outside per week, and weighted lifetime average latitude of residence. Latitude of residence was defined by the absolute value of the latitude from the equator (eg, 10° latitude south and 10° latitude north were treated the same). We performed statistical tests for heterogeneity to assess whether it was appropriate to pool site-specific results.22
We recruited 118 XFS cases and 106 control participants in the United States and 67 XFS cases and 72 control participants in Israel—all were self-reported white individuals (Table 1). Quiz Ref IDAt both sites, cases were older than control individuals, had a higher frequency of a family history of glaucoma, and had light-colored irides (Table 1). While fewer cases had diabetes mellitus at the US site, the opposite was true at the Israeli site. At both sites, cases lived farther from the equator and spent more time outdoors in summer than control participants. A smaller percentage of cases reported ever wearing sunglasses compared with control participants at both sites. Fewer cases compared with control participants wore brimmed hats at the US site but the opposite was true at the Israeli site. At both sites, a higher percentage of cases than control individuals reported ever spending leisure time over water or snow. At the US site, more cases than control participants spent any time working over snow or water but few people at the Israeli site reported these activities.
To explore the interrelationships between the main exposure of lifetime latitude and other potential confounders, we assessed the distribution of covariates in control participants by greater than or less than the median latitude of each cohort (Table 2). At both sites, compared with those residing close to the equator, a greater percentage of people farther from the equator reported any sunburns (70% vs 52% at the US site, P = .07; 65% vs 60% at the Israeli site, P = .57) and spending more time outside (16.2 vs 15.6 h/wk in the US site, P = .98; 15.0 vs 13.0 h/wk in the Israeli site, P = .20). People farther from the equator also wore brimmed hats a greater percentage of the time than those closer to the equator (22.1% vs 14.5% at the US site, P = .22; 21.4% vs 13.5% at the Israeli site, P = .09). More people residing farther from the equator than those living closer wore sunglasses a higher percentage of the time at the US site (53.6% vs 42.5%; P = .27), although there was little difference at the Israeli site (31.7% vs 32.2%; P = .85).
Quiz Ref IDIn pooled analyses (Table 3), after adjusting for age, sex, iris color, family history of glaucoma, diabetes mellitus, and systemic hypertension, each 1° in weighted average lifetime residential latitude away from the equator was associated with 11% increased odds of XFS (pooled odds ratio [OR], 1.11; 95% CI, 1.05-1.17; P < .001). Furthermore, every extra hour per week spent outdoors, averaged over ages 10 to 59 years, was associated with 4% increased odds of XFS (pooled OR, 1.04; 95% CI, 1.00-1.07; P = .03; Table 3). In secondary analysis, we calculated weighted average residential latitude for 3 age periods: 0 to 20, 21 to 40, and 41 to 59 years. In pooled multivariable analysis, average weighted residential latitude during young adulthood (age 21-40 years) was independently associated with XFS (pooled OR, 1.17; 95% CI, 1.05-1.31; P = .01; Table 4), even after adjusting for residence at other periods and other confounders. A similar analysis of time spent outdoors during various age periods in relation to XFS was inadequately powered and did not reveal any associations (data not shown).
In a separate model, the association between sunglass wear and XFS was different between the 2 sites (P for heterogeneity = .005). At the US site, for every percentage-point increase in the lifetime average number of hours sunglasses were worn in the summer between 10 am and 4 pm, the odds of XFS decreased by 2% (OR, 0.98; 95% CI, 0.97-0.99; P < .001); however, no association was observed at the Israeli site (OR, 1.00; 95% CI, 0.99-1.01; P = .92). An association between brimmed hat or visor wear and XFS was not identified (pooled P ≥ .57; models not shown).
Leisure time spent over water or snow was not related to XFS in the US site (data not shown). However, in the US site, any history of work activity over water or snow was associated with nearly 4-fold increased odds of XFS (OR, 3.86; 95% CI, 1.36-10.93; P = .01) after further adjustment for weighted average lifetime residential latitude, sunglass wear, average lifetime hours spent outdoors, and brimmed hat wear. In Israel, there were too few participants who worked or spent leisure time over water or snow for meaningful analysis.
Quiz Ref IDIn our US and Israeli case-control groups, we observed a positive association between residential latitude away from the equator and XFS. This association was driven largely by the varied residential histories of the Israeli participants who lived throughout Europe and in countries of the southern hemisphere. More time spent outdoors in the summer over a lifetime was associated with XFS. Brimmed hat wear was not associated with XFS, while at the US site, sunglass wear was associated with reduced odds of XFS. Taken with the evidence that work over water or snow in the United States was associated with increased odds of XFS, our data suggest that ocular exposure to light from reflective surfaces is important for XFS development.
In the Reykjavik Eye Study, time spent outdoors was not associated with an increased risk for XFS20; however, in that study, only the most recent exposures were considered, which does not necessarily correlate with critical earlier exposures. In youth, people spend more time outdoors23 and have larger pupils,24 making them vulnerable to anterior-segment UV damage. Our survey asked participants to recall their exposures at every decade from ages 10 to 59 years, and we observed associations with greater time spent outdoors averaged over a lifetime.
The amount of UVR incident on the horizontal terrain relates to the ocular UVR dosing in a complicated manner. Using a UV-sensing contact lens, Sydenham et al25 reported that the ratio of ocular to ambient UV exposure ranged from 4% to 23%, indicating that ambient UV exposure is not the major source of ocular solar exposure. Indeed, in our study, brimmed hat wear, which may protect against ambient UV exposure but not radiation reflected from the ground and into the eye, was not associated with XFS. Fresh snow reflects as much as 80% of UV-B (290-320 nm) during midday,26 while sand is an intermediate reflecting surface (approximately 7%-18%) and grass is a poor reflector of UVR (approximately 2%-4%). At the US site, working over snow or water was associated with nearly 4-fold increased odds of XFS, even after controlling for important potentially confounding factors including residential latitude. In our study, sunglass wear at the US site, but not the Israeli site, was associated with reduced odds of XFS. The reason for the difference between the sites is unclear but the data suggest sunglass wear could be associated with reduced odds of XFS by virtue of blocking reflected UVR. The importance of reflected UVR in relation to XFS is further supported by the relatively high prevalence of XFS in Saudi Arabia (9%), where there is considerable sand and sun exposure.27
If UVR is important in XFS as our study suggests, then XFS could be considered as an ophthalmoheliosis (solar-related ophthalmic condition) and the association between other ophthalmohelioses in relation to XFS warrants study to further test the UVR-XFS hypothesis. Climatic droplet keratopathy is a classic ophthalmoheliosis linked to XFS in several studies.1,2,28,29 Pterygium is another ophthalmoheliosis related to UVR that typically presents at a much younger age than XFS.30 While Taylor13 found a connection between pterygium and XFS in aboriginal Australians, this association is not well documented in other populations. Pterygia are thought to result from the Coroneo effect,31 where horizontal UVR rays reflect across the corneal dome to focus on the nasal limbus, and not from light rays directly entering the iris, an important focus of UVR-related pathophysiology in XFS. Thus, the association between ptyergia and XFS may not be particularly strong. Cataract is associated with UV-B exposure32- 34 and has also been associated with XFS.35- 40
One needs to reconcile the role UVR plays in XFS with the positive relation between latitude and XFS.17,18 During the summer, UV doses to the horizon at higher latitude is comparable with that recorded at lower latitude. For example, Godar et al41 estimated that the terrestrial erythema weighted solar UV dose during summer in Boston, Massachusetts (latitude, 42.4° north), was 272 113 J/m2 and 296 436 J/m2 in Atlanta, Georgia (latitude, 33.7° north). This fact, coupled with the higher snow accumulation, which may reflect UVR into the eye during winter months, may explain why Stein et al17 found that Massachusetts had higher hazard ratios (HRs) for XFS (HR, 3.2) relative to Missouri, the geographical center of the United States. In our analysis of the relation between the time spent outdoors and XFS, we measured participant exposure during the summer months when UVR is highest (10 am to 4 pm)26,42 and we controlled for residential latitude.
Quiz Ref IDIt is important to also consider the relatively low prevalence of XFS in Greenland Inuits and Peruvian people,19,43 despite their exposure to considerable reflected sunlight from snow that could contribute to XFS. In addition, while pterygium is a common problem in China,44 XFS is relatively uncommon.45 It should be noted that while the frequency of LOXL1 gene variants related to XFS are unknown among Inuits and Peruvians, these polymorphisms are common in China.46,47 We speculate that Inuits and Peruvians, like Chinese people, have relatively thick irides that may ameliorate uveal tract damage caused by the expected high degree of reflected UVR.48
Several limitations of this study should be considered. The questionnaires used to assess ocular UV were subjective and imprecise. Also, they are vulnerable to inaccurate recall of exposures from the distant past, although UV exposure during young adulthood may be important in XFS. Interviewer bias cannot be ruled out but was minimized with the use of trained interviewers who were masked to ophthalmic status. Another source of bias in our study was that XFS cases may differentially recall more UV exposure than control individuals but it is unlikely that cases would equate working over snow with XFS or appreciate that brimmed hats might be ineffective at blocking reflected UV light from entering the eye. Our XFS cases may have had more UV exposure than typical XFS cases and our control participants may have had less UV exposure than individuals drawn from the general population. However, our data showing an association between the time spent outdoors and XFS in 2 sites is consistent with data showing that indigenous people, who generally spend considerable time outdoors, have higher rates of XFS than people living in comparable urban settings. For example, the prevalence of XFS in Navajo Indians in Arizona was 6%49 compared with clinic-based estimates from southeastern United States of 1.4% to 3%.50,51
This work provides evidence for a role of reflected UV rays in contributing to XFS. It by no means excludes other genetic and environmental mechanisms in XFS pathogenesis. If confirmed in other studies, there could be reason to consider more widespread use of UV-blocking eyewear in the prevention of XFS.
Corresponding Author: Louis R. Pasquale, MD, Department of Ophthalmology, Harvard Medical School, Massachusetts Eye and Ear Infirmary, 243 Charles St, Boston, MA 02114 (firstname.lastname@example.org).
Submitted for Publication: March 25, 2014; final revision received May 25, 2014; accepted July 14, 2014.
Published Online: September 4, 2014. doi:10.1001/jamaophthalmol.2014.3326.
Author Contributions: Dr Pasquale had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Pasquale, Jiwani, Rhee, Levkovitch-Verbin.
Acquisition, analysis, or interpretation of data: Jiwani, Zehavi-Dorin, Majd, Rhee, T. Chen, Turalba, Shen, Brauner, Grosskreutz, Gardiner, S. Chen, Borboli-Gerogiannis, Greenstein, Chang, Ritch, Loomis, Kang, Wiggs, Levkovitch-Verbin.
Drafting of the manuscript: Pasquale, Majd, Kang.
Critical revision of the manuscript for important intellectual content: Pasquale, Jiwani, Zehavi-Dorin, Rhee, T. Chen, Turalba, Shen, Brauner, Grosskreutz, Gardiner, S. Chen, Borboli-Gerogiannis, Greenstein, Chang, Ritch, Loomis, Wiggs, Levkovitch-Verbin.
Statistical analysis: Jiwani, Majd, Loomis, Kang.
Obtained funding: Pasquale, Jiwani, Wiggs.
Administrative, technical, or material support: Pasquale, T. Chen, Turalba, Shen, Grosskreutz, Levkovitch-Verbin.
Study supervision: Pasquale, Rhee, Gardiner, Levkovitch-Verbin.
Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Dr Ritch has received personal fees from iSonic Medical, Sensimed, Aeon Astron, Taejoon Pharmaceutical, Pfizer, Merck, Allergan, and Ocular Instruments Inc outside this work. Dr Wiggs has served on the scientific advisory boards for the Glaucoma Research Foundation and Research to Prevent Blindness. No other disclosures were reported.
Funding/Support: This work was supported by the Arthur Ashley Foundation, grant EY020928 from the National Institutes of Health (Dr Wiggs), a Physician Scientist Award from Research to Prevent Blindness (Dr Pasquale), a Harvard Medical School Ophthalmology Scholar Award (Dr Pasquale), and a Doris Duke Charitable Foundation grant (Dr Jiwani). Dr Wiggs has received an honorarium for travel expenses from Research to Prevent Blindness.
Role of the Funder/Sponsor: The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.