Cruickshanks KJ, Klein R, Klein BEK, Nondahl DM. Sunlight and the 5-Year Incidence of Early Age-Related MaculopathyThe Beaver Dam Eye Study. Arch Ophthalmol. 2001;119(2):246-250. doi:10-1001/pubs.Ophthalmol.-ISSN-0003-9950-119-2-eeb00005
To investigate the relation of sunlight exposure and indicators of sun sensitivity with the 5-year incidence of early age-related maculopathy (ARM).
Longitudinal, population-based study. Participants (aged 43-86 years at baseline) in the Beaver Dam Eye Study were reexamined from 1993 to 1995, 5 years after the baseline examination. Questionnaire data about sunlight exposure and sun sensitivity were obtained at baseline. Additional information about earlier life patterns of exposure was ascertained at follow-up. Stereoscopic color fundus photographs were graded to determine the presence of ARM at the 5-year follow-up in eyes free from signs of early ARM at the baseline examination.
Leisure time spent outdoors while persons were teenagers (aged 13-19 years) and in their 30s (aged 30-39 years) was significantly associated with the risk of early ARM (odds ratio, 2.09; 95% confidence interval, 1.19-3.65). There was a slight, but nonsignificant, protective effect associated with use of hats and sunglasses while persons were teenagers and in their 30s (odds ratio, 0.72; 95% confidence interval, 0.50-1.03). People with red or blond hair were slightly more likely to develop early ARM than people with darker hair (odds ratio, 1.33; 95% confidence interval, 0.97-1.83). There were no associations between estimated ambient UV-B exposure or markers of sun sensitivity and the incidence of early ARM.
Exposure to sunlight may be associated with the development of early ARM.
IDENTIFYING primary prevention strategies for age-related maculopathy(ARM) is an important goal for epidemiologic research as there are no effective sight-saving therapies for most patients with this vision-threatening disease. There is growing evidence that genes1- 3 and environmental factors4- 9(including lifestyle factors such as smoking and other chronic diseases such as atherosclerosis) may be important in the multifactorial etiology of ARM. Thus far, however, much of the epidemiologic evidence for environmental influences has been from cross-sectional cohort and case-control studies. Ocular exposure to sunlight (blue light) has been associated with ARM in a cohort of men with high levels of sunlight exposure during work (Maryland watermen).5 In the Beaver Dam Eye Study,9 people who reported spending most of their leisure time outdoors in the summer were more likely to have late maculopathy (odds ratio [OR], 2.19; 95% confidence interval, 1.12-4.25). Longitudinal studies are needed to investigate this association and strengthen the evidence for a causal relation. The present study describes the relations between reported sunlight exposure and the 5-year incidence of early ARM in the Beaver Dam cohort.
The Beaver Dam Eye Study is a population-based study of age-related ocular disorders. Details of the methods have been published previously.10- 14 A private census was conducted to identify all persons between the ages of 43 and 84 years who were residents of the city or township of Beaver Dam, Wis, from 1987 to 1988.10 Of the 5925 eligible persons identified, 4926 (83.1%) were examined during the baseline examination(March 1, 1988, through September 15, 1990), 225 (3.8%) died before examination, 91 (1.5%) had moved out of the area, 23 (0.4%) could not be located, 269 (4.5%) completed a questionnaire only, and 391 (6.6%) refused to participate (percentages may not total 100 because of rounding). Nonparticipants were older than participants.
From 1993 to 1995, a 5-year follow-up examination of the cohort was conducted using the same methods as used in the first examination.12,14 Of the 4541 surviving baseline participants, 3684 (81.1%) participated in the follow-up examination, 423 (9.3%) refused to participate, 259 (5.7%) completed a questionnaire only, 171 (3.8%) died before examination, and 4 (0.1%) could not be located. Surviving participants who were not reexamined in the follow-up eye study (n = 686) were older, had less education, had a lower income, had poorer visual acuity, were more likely to have a history of cardiovascular disease, and smoked more than participants in the 5-year follow-up examination.14 After adjusting for age and sex, there were no significant differences between nonparticipants and participants in baseline sunlight exposure variables (data not shown).
During each examination, a medical history questionnaire was administered. Participants were asked about residential history; time spent outdoors during leisure and work; and use of eyeglasses for distance vision, hats with brims, and sunglasses.9 At the 5-year follow-up examination, additional questions were added to ascertain sunlight-related behaviors during their teenage years (age 13-19 years) and their 30s (age 30-39 years).
An index of ambient UV-B exposure was constructed from the baseline residential history, weighting time spent outside of Wisconsin by the ratio of the total ambient UV-B light present in that area to the level for 1 year in Wisconsin (Wisconsin sun year).9 The average annual ambient UV-B exposure was calculated by dividing the cumulative ambient exposure by age. Most participants had spent most of their lives in Wisconsin, resulting in a highly skewed distribution of average annual exposure, so this variable was used to categorize participants into 2 levels of exposure. Between the baseline and the follow-up examinations, some people had migrated out of Beaver Dam to other regions with higher UV-B levels or had spent vacation time in areas with higher UV-B levels. Therefore, an additional index was created using the same weighting procedure as used for the historical data, to capture additional exposure during follow-up to areas with higher UV-B levels than Wisconsin.
The amounts of time participants reported using hats and sunglasses at the baseline examination were combined in a weighted fashion into levels of increasing protection from UV-B light (none, low, moderate, and high). A high protection level was defined as wearing hats, sunglasses, or both at least half of the time spent outdoors.9
Indexes of sunlight exposure and use of hats and sunglasses while aged 13 to 19 years and 30 to 39 years were created. Participants who reported being outside in summer for 5 or more hours per day in both age periods were considered to have high exposure, those who reported being outside less than 2 hours per day in both age periods were considered to have low exposure, and those with 2 to 5 hours of exposure per day or exposure that varied by age period were considered to have intermediate exposure. Similarly, participants who reported they wore hats or sunglasses at least half the time while outside in the summer in both age periods were considered to have used them often and have a "high" level of protection. People who reported "rarely" using either hats or sunglasses in both age periods were considered to have a low level of protection. Participants who reported variable patterns of use were considered to have intermediate levels of protection.
Age-related maculopathy was assessed by grading stereoscopic 30° color fundus photographs taken at the baseline and 5-year follow-up examinations using the Wisconsin Age-Related Maculopathy Grading System.13- 15 Stereoscopic photographs were taken centered on the disc (Diabetic Retinopathy Study standard field 1) and macula (Diabetic Retinopathy Study standard field 2), and a nonstereoscopic color fundus photograph was taken temporal to but including the fovea of each eye.
Among eyes free from ARM at baseline, the incidence of early ARM was defined as the presence, at follow-up, of soft indistinct drusen or any type of drusen associated with either retinal pigment epithelial depigmentation or increased retinal pigment. The 5-year incidence of late ARM was low (0.9%), so this end point was excluded from further analyses.14
SAS statistical software, version 6.09 (SAS Institute Inc, Cary, NC), was used to calculate the χ2 statistic to test for overall associations, to test for trends in proportions, and to determine logistic regression models. Binary logistic regression models were used to examine the relations of sun-related variables with the incidence of early ARM, controlling for the effects of age and other covariates. Sun-related exposure variables were also entered into logistic regression models in combinations to test for associations while controlling for the potentially important modifying effects of other exposure measures.
Overall, the incidence of early ARM was 8.2% (Table 1). The incidence was higher for older vs younger age groups. In univariate analyses (Table 2), the 5-year incidence of early ARM was significantly higher among those who reported having blond or red hair than among people with brown or black hair. The incidence of early ARM was lower among those who reported, at the time of the baseline examination (1988-1990), usually using hats and sunglasses while outside. A similar pattern was found for use of these protective devices in earlier age periods. The incidence of early ARM was 6.9% for frequent users of hats, sunglasses, or both while persons were teenagers and in their 30s compared with 9.8% for those who rarely used either hats or sunglasses.
The age- and sex-adjusted ORs are given in Table 3. Leisure time spent outdoors in summer while persons were aged 13 to 19 and 30 to 39 years was significantly associated with the incidence of early ARM. There were slight, but nonsignificant, associations (P ≥ .05) with the use of hats and sunglasses during the same age periods and with having blond or red hair compared with dark hair. None of the other sunlight exposure variables or sun sensitivity variables were associated with the incidence of early ARM. There was no association with the marker of ambient UV-B exposure based on residential history at baseline and the 5-year incidence of early ARM. Including a measure of ambient UV-B exposure during the 5-year follow-up period did not alter this finding (data not shown).
Additional age- and sex-adjusted models testing the effects of combinations of the sun-related variables (exposure and sensitivity) were evaluated, and interaction terms were included as appropriate. The final model retained, in addition to age and sex, leisure time spent outdoors in the summer while persons were teenagers and in their 30s, use of hats and sunglasses during those periods, and hair color (Table 4). While the latter 2 variables were not statistically significantly associated with the incidence of early ARM, they did modify the strength of the association between time spent outdoors and incidence. Participants who spent 5 or more hours per day outside during leisure time when they were aged 13 to 19 and 30 to 39 years were twice as likely to develop early ARM as those who reported spending less than 2 hours a day outside.
Cigarette smoking has been demonstrated, in the Beaver Dam cohort, to be associated with the incidence of ARM.6 Therefore, an additional model was run, adding indicator variables for current and past smoking (vs never smoking). Leisure time spent outdoors while persons were teenagers and in their 30s remained significantly associated with the risk of early ARM (OR, 2.09; 95% confidence interval, 1.19-3.65). Adding additional lifestyle variables of beer drinking and vitamin use did not alter this association(data not shown).
In this study, people who reported spending 5 or more hours per day outside in the summertime during their teenage years and 30s appeared to have a greater risk of developing early ARM as older adults than people who spent little time outside. There was a slight attenuating effect of wearing hats and sunglasses during these age periods. These data are consistent with the cross-sectional epidemiologic studies5,9 that found that extended exposure to bright light may be associated with ARM. Animal studies16- 19 also have suggested that exposure to bright sunlight can induce retinal changes. Among humans, retinal damage has been found with sun gazing and prolonged exposure to intense, bright light at the beach or in the desert.20 While the incidence of late ARM was too low to permit evaluating the link with the development of the more severe, vision-threatening forms of ARM, this is the first prospective study, to our knowledge, to find a link between sunlight exposure and risk of early ARM.
There was no association between UV radiation exposure and early ARM incidence in this cohort. To date, in spite of experimental studies that suggest that the retina is susceptible to UV-B damage, there is no evidence from epidemiologic studies of any risk, suggesting that the filtering effects of the lens provide ample protection throughout the life course.21 Interestingly, aphakic individuals in Beaver Dam were more likely to develop late ARM.22
There were no statistically significant associations between baseline sunlight exposure variables and the 5-year incidence of early ARM, which may reflect the smaller number of older adults still spending time outdoors in summer and winter. In fact, the point estimate suggests a weak link between time spent outside in winter at baseline and the incidence of early ARM.
The slight, but not significant, excess risk for blond- or red-haired people may reflect their lighter pigmentation and, therefore, less protection from the damaging effects of light, which may suggest some genetic predisposition that clusters within certain ethnic backgrounds, or may represent a chance finding. Vinding,23 in a cross-sectional study, found no association between hair color and risk of late ARM.
The Blue Mountains Eye Study24 recently reported an association between abnormal skin sensitivity to sunlight (either increased or decreased) and ARM. We found no association between skin sensitivity and early ARM. Reasons for this discrepancy are not clear, but may relate to the differences in amount of exposure to sunlight in the Australian population compared with this population, who have spent most of their lives in a northern region of the United States, where opportunities to tan or burn are few. A recent case-control study25 in Australia, which measured ocular exposure to sunlight by retrospective questionnaire, found that among poor tanners, the average exposure was more than 600 hours per year.
In our study, only retrospective information about time spent outside during 2 earlier periods was available. Clearly, long ago habits may be poorly and inaccurately recalled by older adults. However, the respondents were asked to broadly classify habits, which may have lessened recall problems. The true impact of possible misclassification bias cannot be assessed. These results may be spurious and reflect uncontrolled confounding, as people who spend major portions of their time outside differ in many ways from those who tend to be indoors. Models including markers of other health habits did not alter the association, but there may have been unmeasured confounders of this relation. Although these data should be viewed with caution, accurate assessment of past sunlight exposure and long-term studies with prospectively measured exposures will remain problematic.
These results fit with a small body of epidemiologic evidence that indicates that exposure to bright sunlight may be a risk factor for early ARM. Furthermore, they support the idea that ARM may be partly preventable through modification of lifestyle factors.6,26,27 Longer follow-up times or larger cohort studies are necessary to have sufficient power to evaluate the link with late-stage ARM, a critical step in determining the etiologic significance of sunlight in the pathogenesis of ARM.
Accepted for publication July 6, 2000.
This study was supported by grant EY06594 from the National Institutes of Health, Bethesda, Md (Drs R. Klein and B. E. K. Klein); and by the Lew R. Wasserman Award from Research to Prevent Blindness Inc, New York, NY (Dr Cruickshanks).
Corresponding author and reprints: Karen J. Cruickshanks, PhD, Department of Ophthalmology and Visual Sciences, University of Wisconsin, 610 N Walnut St, Room 460, WARF Building, Madison, WI 53705-2397 (e-mail: firstname.lastname@example.org).