|Home | About | Journals | Submit | Contact Us | Français|
The relationships between lifestyle behaviors of diet, smoking and physical activity and the subsequent prevalence of age-related macular degeneration (AMD) were investigated.
The population included 1,313 participants (55 to 74 years) in the Carotenoids in Age-Related Eye Disease Study (CAREDS), an ancillary study of the Women's Health Initiative Observational Study (WHIOS). Scores on a modified 2005 Healthy Eating Index (mHEI) were assigned using responses to a food frequency questionnaire administered at WHIOS baseline (1994-1998). Physical activity and lifetime smoking history were queried. An average of six years later, stereoscopic fundus photographs were taken to assess presence and severity of AMD; present in 202 women, 94% of whom had early AMD, the primary outcome.
In multivariate models, women whose diets scored in the highest compared with the lowest quintile on the mHEI had a 46% lower odds for early AMD. Women in the highest vs. lowest quintile for physical activity (MET- Hrs/Wk) had 54% lower odds for early AMD. Although smoking, alone was not independently associated with AMD, having a combination of three healthy lifestyles (healthy diet, physical activity and not smoking) was associated with a 71% lower odds for AMD compared with having high risk scores (P=0.0004).
Modifying lifestyles might reduce risk for early AMD as much as 3-fold, lowering the risk for advanced AMD in a person's lifetime and the social and economic costs of AMD to society.
The macula of the human eye progressively degenerates with age, more quickly in some people than in others, and can lead to advanced age-related macular degeneration (AMD), which involves the loss of photoreceptors in the macula of the eye. Treatment for advanced AMD is of limited effectiveness, is a costly disease to treat 1 and will become more so as the number of older Americans increase in the coming decades.2 Moreover, it profoundly limits the ability of older adults to function independently. The loss of central vision associated with advanced AMD diminishes the ability to see and recognize other people's faces, and to read fine print such as in newspapers and on pill bottles and food packages.
Early and advanced stages of AMD are consistently more common in people who have specific genotypes (many of which function to mediate response to inflammation, oxidative stress, lipid metabolism and angiogenesis, as recently reviewed. 3) Several modifiable aspects of lifestyle have been related to a lower occurrence of AMD, including smoking, physical activity and diet. AMD is also sometimes observed to be more common in people with a history of chronic diseases or conditions which can also be modified by lifestyle choices 4-7 such as cardiovascular disease, 5, 6 diabetes,7 hypertension, 2,8 obesity, 9-11 and diseases of inflammation or elevated markers of inflammation.12
Smoking has been the most consistently reported risk factor for AMD.13 However, these associations may reflect other unhealthy lifestyle habits which are more common in smokers; the associations of smoking to AMD are not often adjusted for other aspects of healthy lifestyles which are less common in smokers. Physical activity has been only recently studied in relation to AMD and was related to lower risk for advanced AMD in three past studies. 14-16
AMD has often been associated with diets which are poor in one or more ways: It has been more common among people with low levels of carotenoids in the diet, serum or macula and diets or serum low in one or more other nutrients, or diets high in fat (reviewed 17). However, associations of single nutrients to AMD are often inconsistent across studies and impossible to totally disentangle from other aspects of diet. Moreover, they do not account for synergistic relationships of food components. Recently, combinations of nutrients 18, 19 or a diet score which reflects 1990 US Dietary Guidelines have been related to lower risk of AMD. 20 Relationships of AMD to two currently recommended diet patterns (The 2005 US Dietary Guidelines or a Mediterranean diet pattern) have not been previously reported.
The common approaches to assessing relationships of healthy lifestyles to AMD in observational studies may give erroneous estimates of single aspects of healthy lifestyles. This is, in part, because single aspects of diet or lifestyle are difficult to disentangle from each other. We cannot measure the levels of these individual aspects of a healthy lifestyle perfectly across several decades of adult life when they are likely to influence AMD; any attempt to simply adjust one in consideration of the other(s) risks the likelihood of incomplete adjustment or residual confounding.
Moreover, adjustment of one healthy lifestyle for another may lead to misleading interpretations. This is because the mechanisms of protection of some healthy lifestyles are related. For example, the energy expenditure of physical activity permits a higher daily nutrient intake and may be protective in this way. Also, both physical activity and diet can contribute to better vitamin D status 21 which has been related to lower risk for AMD. 22
If several aspects of lifestyle all protect through a common mechanism (such as reducing inflammation) then examination of risk associated with one healthy lifestyle can be underestimated. Single studies have not previously considered these risk factors, together with diet, concurrently. The objective of the present report was to describe relationships of AMD to a combination of healthy behaviors, including diet, physical activity and smoking history. This was possible in the Carotenoids in Age-Related Eye Disease Study because participants were recruited from a sample of women who provided detailed dietary and lifestyle habit histories as part of the Women's Health Initiative (WHI) an average of six years before AMD was assessed.
Women (50 to 79 years of age) were recruited into the CAREDS from those who were enrolled in the Women's Health Initiative Observational Study Cohort (WHI-OS)23 at 3 of 40 sites: the University of Wisconsin (Madison), the University of Iowa (Iowa City), and the Kaiser Center for Health Research (Portland). Women who had intakes of lutein plus zeaxanthin that were above the 78th and below the 28th percentiles (n=3143), as assessed at WHI baseline (1994-1998) were sent letters inviting them to participate in the eye study. Sampling women at the extremes of dietary intake maximized the statistical power available to detect associations between AMD and levels of lutein and zeaxanthin in the diet and serum. Because the intake of lutein and zeaxanthin is also correlated with intakes of many vitamins and minerals from foods (range of Spearman correlation coefficients = 0.27 for vitamin D to 0.77 for folate) and negatively correlated with fat intake (r = -0.36), this design would also be expected to maximize extremes in intake of other aspects of healthy diets and enhance the power available to detect associations with these related aspects of diet, relative to samples with similar ranges of intake of comparable sizes.
Of the 3,143 women recruited, 2,005 (64%) were enrolled and photographic evidence of AMD was determined in CAREDS examinations in 2001-2004, four to seven years (mean of 6.3 years) after WHIOS baseline. Gradable fundus eye photography was completed on 1,853 eligible women, of whom 1,787 provided full detail regarding covariates used in regression models of AMD. Based on evidence for selective mortality bias in associations of diet to AMD in women over 75 years of age,24, 25 and similar findings in relation to the independent variables which are evaluated in the present analysis (data not shown), the present dataset includes only the 75% of women in this sample who were less than 75 years of age at the time of eye photography (N=1,325 women). All procedures conformed to the Declaration of Helsinki and were approved by the Institutional Review Board at each University.
A comparison of CAREDS participants and non-respondents in the full dataset has been previously described.24, 25 Further, we compared CAREDS participants < 75 years (N= 1325) with WHIOS participants the same ages who were recruited, but did not participate, or were excluded from our analysis dataset because of missing covariate data (n=922). Four percent of women in this analysis reported having physician diagnosed AMD at WHIOS 3 year follow-up visits vs. 2 percent of the non- participating women (p=0.13). Women in this analysis were (after age-adjustment) more likely to have never smoked (53 vs. 49%; p=0.0002), and had diets that were slightly lower in fat (32 vs. 33% of energy; p = 0.001) and higher in lutein and zeaxanthin (1.8 vs. 1.6 mg/day; p <0.0001), reported higher levels of physical activity (15 vs.12 MET-Hr-Wk; P<0.0001), and had lower body mass index (BMI) (median =27.7 vs.28.4 Kg/M2).
Measures of macular pigment density26 and dilated fundus photography27 taken at CAREDS baseline (2001-2004) have been described. Iris color was determined from retina photographs. Stereoscopic fundus photographs were graded for AMD by the University of Wisconsin Fundus Photography Reading Center, using methods based on those used in the Age-Related Eye Disease Study (AREDS).27
The primary outcome was the presence of early AMD in at least one eye. Early AMD was defined as the presence of either: large drusen (one or more large drusen (≥ 125μm) or extensive intermediate drusen (area ≥ 360μm when soft indistinct drusen is present; or, ≥ 650μm when soft indistinct drusen is absent) or pigmentary abnormalities of the retinal pigment epithelium (an increase or decrease in pigmentation accompanied with at least one druse (≥ 63μm)), consistent with previously established definitions. 28 This corresponds to stage 3 of the AREDS original AMD definitions, with the exception that, like other population-based studies, it includes, in the definition of early AMD, having pigmentary abnormalites (with drusen). All analyses were also performed separately for two components of early AMD (large drusen and pigmentary abnormalities).
There were only twelve women (n=12) with advanced AMD in the CAREDS sample <75 years of age (defined as geographic atrophy, neovascularization, or exudation in the center sub-field or receiving a physician diagnosis of advanced AMD, confirmed subsequently in writing by a physician.) Because this was too small of a number to analyze this endpoint separately, and to reduce the possibility of temporal bias influencing the estimates of relationships of healthy lifestyles to early AMD, we excluded these women from the main analyses.
Daily levels of nutrients in diets were estimated from responses to a previously validated, semi-quantitative food frequency questionnaire (FFQ)29 at WHI baseline. For this report, we primarily evaluated diets using the 2005 Healthy Eating Index (HEI-2005) which reflects adherence to the 2005 US Dietary Guidelines.30
The HEI-2005 assigns scores for the intake of specific food components (per 1000 kilocalories), as shown in Table 1. This score represents different aspects of a healthy diet, including the abundant intake of fruits and vegetables (including those which are particularly nutrient-rich) and whole grains, and low intake of discretionary calories from sugar, fat and alcohol and of saturated fat. We made one modification to the original 100-point HEI-2005 scoring system, herein referred to as the modified HEI-2005 (mHEI), by omitting the 10 points assigned for the intake of oils (non hydrogenated vegetable oils and oils in fish, nuts and seeds). The original intent for including points for oils in the scoring system was to ensure adequate intake of vitamin E and essential fatty acids. However, using this component, in evaluation of diets from Americans who otherwise consume adequate sources of these nutrients, may lead to higher scores for diets which are less nutrient dense. Indeed, omitting this score component (which largely reflected vegetable oil intake in this sample) improved correlations with several vitamins and minerals, while not influencing associations with blood levels of alpha-tocopherol (data not shown).
We also explored whether a similar association of AMD to another healthy diet pattern existed by assigning scores for adherence to a Mediterranean diet adapted for use in American people (aMED).31 In a nine-point scoring system, a point was assigned for 1) servings of each of the following food components >75th percentile within the sample: fruits, vegetables, whole grains, legumes, nuts and fish, and ratio of monosaturated to saturated fat and; 2) <25th percentile for servings of red meat, and 3) alcohol intake of 5-25 g/day. These scores were associated with a wider distribution of intakes of many vitamins and minerals than the mHEI (data not shown).
While results are reported by both scoring systems, for the opportunity of greater insights, we focus on the mHEI and used the aMED scoring system for further exploratory analyses. This is because scores are spread more widely on the mHEI score (90 possible points) than the aMed score (9 possible points). Moreover, high scores on the aMED pattern were uncommon in this sample; only 53 women had scores of 6-9. The mHEI has the added advantages of being based on established recommendations for reduction of chronic disease risk in Americans,32 and being easier for the comparison of results across American study samples.
At WHIOS baseline women were asked about participation in total and recreational physical activity which included household and yard work, walking and strenuous, moderate, and intensive activities.33 Responses to these questions were used to estimate energy expenditure in metabolic energy task (MET)-Hrs/wk. These reflect the sum of METs multiplied by the duration and frequency of activity in a week. MET values are based on estimates of the intensity of the physical activity; for example, one hour of strenuous activity requires 7 METs, one hour of moderate activity or walking very fast requires 4-4.5 METs and one hour of low intensity or walking slow requires 3 METs.
At WHIOS baseline women were weighed and measured and their body mass indexes (BMIs) (Kg/M 2) were computed and blood pressure measured. At the time of entry into the WHIOS women provided information about their smoking history. This was updated in CAREDS questionnaires, permitting the creation of a summary variable of lifetime smoking (pack years), further categorized as 1) never smoked, 2) smoked 0-7 pack years or 3) smoked > 7 pack years. Additional demographic, lifestyle, and health history data were available from questionnaires including: education, hormone replacement therapy use, alcohol use, pulse pressure, and history of diabetes, hypertension, and cardiovascular disease. At CAREDS visits, we also queried family history of macular degeneration (immediate family member age 65 years or older when diagnosed) and obtained updated histories of alcohol use and diabetes. WHI baseline serum samples, collected after a ≥ 10-hour fast, were stored at -80 degrees centigrade and analyzed in 2004-2005 for serum levels of lutein and zeaxanthin and tocopherols by a reverse phase HPLC analysis,34 and in 2007 and 2008 for serum 25- hydroxyvitamin D by the Diasorin LIAISON® chemiluminescence method and high sensitivity C-reactive protein (CRP) by High sensitivity CRP Assay kit (DiaSorin, Stillwater MN) at Heartland Assays, Inc. (Ames, IA).
We constructed a 6-point healthy lifestyle score (HLS) which gave equal weight to three levels of each of three health habits: diet, physical activity and smoking, based on our knowledge of the distribution of these variables. We assigned 2 possible points for healthy levels of each behavior: diet (mHEI score in lowest 20 percent= 0, 21- 80th percent= 1 and in the highest 20 percent = 2), physical activity (MET hrs/Wk) (lowest tertile= 0, second tertile = 1 and third tertile = 2) and smoking (>7 pack years = 0, >0-≤7 = 1 and none/never smoking = 2).
We evaluated the relationships of quintile ranks for scores on mHEI and physical activity to other risk and protective factors for AMD using ANCOVA and Cochran-Mantel-Haenszel. Next, logistic regression (PROC LOGISTIC in SAS v.9.2, SAS Institute, Cary, NC) was used to compute odds ratios (OR) and 95 percent confidence intervals (95% CI) for specific AMD endpoints (early AMD, large drusen, pigmentary abnormalities, excluding women with advanced AMD from the reference groups (N=12), and any AMD). OR were described by quintile or categories, as shown in Table 3 and P values for trend across the continuous range of score values was also computed.
OR for AMD were first adjusted for age and multivariate models were further adjusted for other risk factors which were not explanatory or intermediary variables. Previous multiple variable models in this sample 24, 25 included age, pack-years smoked, history of diabetes, family history of AMD, iris color, history of cardiovascular disease, and use of hormone replacement therapy. These were included in the multivariate-adjusted models unless they were independent variables. Additional variables which were statistically significantly related (p≤0.10) to both AMD and healthy diet pattern or healthy lifestyle scores in CAREDS, or which were previously suspected to be biologically plausible confounders, were tested by adding them singly to age-adjusted models. (For mHEI score only non dietary-risk factors were tested.) Those covariates that changed the AMD odds ratio by ≥10% were retained in the final model. We tested for interactions to determine whether mHEI or physical activity associations with AMD differed (p for interaction <0.10) by age, study site, family history of AMD or levels of smoking or BMI.
Women whose mHEI diet scores were ranked in the highest, compared with the lowest quintile, had lower rates of early AMD (11% vs. 19%), (Table 2), diets significantly lower in fat (% energy) and higher in median servings of several food groups (fruits, vegetables, dairy, grains and meats or alternatives (including poultry, meat, fish, beans and eggs) (Table 3). Supplement use was fairly common: 56 vs. 37% of women in high, compared with low quintiles for mHEI used multivitamins. Being in high vs. low quintiles for mHEI score was also associated with reporting more physical activity, fewer years of smoking, a lower likelihood of having a history of hypertension, lower systolic blood pressure, and lower levels of BMI and serum CRP (Table 3. Associations with aMED scores and non nutritional AMD risk factors were generally similar (data not shown).
Women in the highest, compared to the lowest, quintile for mHEI score had 46% lowering of odds for early AMD multivariate-adjusted OR (95% CI) = 0.54 (0.33-0.88) (Table 4). The 58 women with aMED scores of 6-9 had a 66% lower odds for AMD, compared with many more women (n=490) in this sample who scored 0 or 1 on this pattern (Table 4). Further adjusting for physical activity attenuated ORs or early AMD in women with high compared to low scores for both diet patterns; Multivariate-adjusted OR (95%CI) for early AMD among women in high vs. low quintiles for mHEI = 0.64 (0.38-1.07) and among women with aMED scores of 6-9 vs. 0-1 = 0.44 (0.10-1.27), indicating that associations with diet is not totally independent from the association with physical activity.
The multivariate-adjusted OR for large drusen and pigmentary abnormalities among women in highest compared to lowest quintiles of the mHEI were similar to those for overall early AMD (multivariate-adjusted OR (95%CI) = 0.56 (0.31-0.97); P for trend = 0.049 and 0.58 (0.29 – 1.13); P for trend = 0.07), respectively. Associations of mHEI to total AMD (early plus advanced AMD) were also similar (multivariate-adjusted OR (95%CI) = 0.55 (033-0.88); P for trend = 0.012).
The associations between mHEI score and early AMD were significantly different across study sites (P for interaction = 0.08), with stronger inverse associations between mHEI score and AMD in Portland (n=425) and Madison (N=436) than in Iowa City (N= 452), but all associations were inverse (data not shown). The inverse associations of mHEI score to AMD did not significantly differ (p > 0.20) by BMI, physical activity level, smoking history, macular pigment density level or having a family history of AMD (data not shown).
Women in the highest quintile vs. lowest quintile for physical activity had a greater than two-fold lower multivariate-adjusted odds for AMD (OR (95%CI) = 0.46 (0.27-0.78); P for trend = 0.002). Associations with drusen, pigmentary abnormalities and any AMD were similar (data not shown). Despite a significant correlation between level of physical activity and mHEI score (Spearman correlation coefficient = 0.30; p<0.0001), adjusting for mHEI only slightly attenuated this association (OR (95%CI) = 0.52 (0.30-0.89); P for trend = 0.009). The association also remained consistently inverse in women with mHEI scores above and below the median (P for interaction = 0.90). Physical activity was also correlated with BMI (r=-0.26; p<0.001). However, the association of physical activity to early AMD was also consistently inverse across all levels of BMI (<25, 25-29 and >30; P-value for interaction = 0.33).
The association of physical activity to early AMD appeared to reflect the weekly time spent in physical activity rather than any specific type or intensity. For example, multivariate-adjusted OR (95%CI) for women in highest vs. lowest tertiles (minutes/week) for total recreational activity were 0.56 (0.37-0.84); P for linear trend=0.004, for moderately strenuous activity were 0.78 (0.52-1.16); P for linear trend = 0.01, and for strenuous activity were 0.67 (0.46-0.96); P for linear trend = 0.004.
The OR for early AMD by smoking history and BMI are also given in Table 4. Women who smoked more than 7 pack-years had a 45 % increase in odds for AMD in multivariate-adjusted models, but the association across all levels of smoking was only marginally significant (P=0.07) and was attenuated after adjusting for mHEI and physical activity. Only the 10% of women who were extremely obese had 58% higher odds for early AMD and after further adjustment for multiple variables, mHEI and physical activity, BMI was no longer associated with AMD (Table 3).
Women who had a HLS of 6, which reflected healthiest (lowest risk levels) of all three score components (diet, physical activity and smoking), had a 71% lower odds for early AMD, compared with women who had scores of 0-2 (Table 4). Obesity (BMI >30) was much less common in women with a HLS of 6 (9%) vs. 0-2 (43%) (P=<0.0001). However, adjusting for BMI did not influence associations. We explored the consistency of the HLS associations with early AMD in women who were obese (BMI ≥30) vs. not obese (BMI < 30). (In these analyses we grouped women with scores of 5-6 because too few women with a HLS of 6 were obese.) The associations were somewhat, but not statistically, stronger in obese women (P for interaction =0.17). The multivariate – adjusted OR (95% CI) were 0. 26 (0.06-0.78); P trend = 0.004 in women who were obese and 0.60 (0.34-1.06); P trend = 0.02 in women who were not.
The results of the present study indicate that broadly healthy diets and lifestyles in women 50-69 years of age were associated with a lower prevalence of early AMD an average of 6 years later. A three-fold lowering of odds was associated with having a combination of healthy lifestyles which included healthy diets, physical activity and not smoking. Specifically, in this particular sample, the 5 % of women with highest lifestyle score (equal to 6) never smoked (<0.1 packyear), reported the equivalent of about 10 hours of low intensity physical activity/d (such as walking or gardening) or 8 hours of moderate activity/week and had the following diet qualities: daily servings of fruits (3.5/day) and vegetables (about 5/day; 2 of which were dark green, orange or legumes), dairy (2.3/day) meat or alternatives (meat, poultry, fish, beans or eggs) (2.7 ounces/day) and grains ( 3.5 servings/day of which 1 serving/day is whole grain). These lifestyle habits are interrelated in practice and in biological effect (discussed below) so that the degree to which they might contribute independently to associations cannot be accurately assessed in this study.
These associations, like those generated from data in prospective studies, are not likely to reflect temporal biases, which are possible in studies in which lifestyle and AMD are assessed at the same point in time. This is because 1) lifestyle was assessed an average of 6 years before photographs documented AMD and 2) most women who had AMD at this point in time were in early stages ( 77% of women who had not previously told they had AMD) and 3) because of their young ages, the women who had AMD were also not likely to have had AMD for many years.
These analyses provide the first estimate of associations between AMD and the dietary patterns 35 recommended by the 2005 U.S. Dietary Guidelines. 32 Associations with mHEI were stronger than associations for individual aspects of diet previously studied in this sample24, 25 Results are consistent with a recently reported association between an alternative version of the 1990 Dietary Guidelines and advanced AMD in a case-control study.20 Further, results extend the protective nature of broadly healthy diet patterns to early AMD which dramatically increases risk for eventually developing advanced AMD in Caucasian populations.36-38
The shift in odds ratio towards unity after adjusting for physical activity suggests some of this effect could be due to physical activity. Both healthy diets and physical activity improve nutritional status. (Recommendations for both are included in the 2005 US Dietary Guidelines.32) Physical activity might contribute to better nutritional health by 1) increasing energy expenditure, allowing a larger absolute intake of phytochemicals and micronutrients and 2) increasing moderate exposure to sunlight, when outdoors, as was common in this sample. Walking outdoors was a strong predictor of blood vitamin D levels in this sample (M. Kluczynski, unpublished manuscript) and a separate sample of WHI participants. 21 Low levels of vitamin D were associated with higher odds for AMD in a sub sample of CAREDS 39 and in a separate sample.22
Having a high score (6-9) on the aMED diet pattern, which is more plant- food focused than the U.S Dietary Guidelines, and predicts somewhat higher intake of several nutrients in the diet and serum than the mHEI score (data not shown), was associated with lower odds for AMD than scores in the top quintile for mHEI (66 vs.46% reduction, respectively in odds). Few women in this sample (4%) had aMED scores in this range. Intake of three foods groups which contribute to higher scores on the aMED, whole grains (Table 3), nuts and fish (previously described25) were limited in this sample. These foods could contribute to the intake of short (grains and nuts) and long-chain (fatty fish from cold water) omega-3 polyunsaturated fatty acids (PUFAs). The intake of fish or omega-3 PUFAs has been associated with lower risk for AMD in many previous studies.40, 41
The protective association of mHEI to AMD in this sample is likely to reflect the fact that high scores were associated with an intake of high levels of several single nutrients which have been related to low prevalence or progression of AMD in previous studies (antioxidants,18, 42 B-vitamins,43 zinc18, 44, 45 and lutein plus zeaxanthin24, 45-49). Previous investigators have found that combinations of nutrients from food are more strongly associated with AMD risk than single nutrients.18, 19 The protective association of mHEI to early AMD is also likely to reflect direct relationships of dietary fat to AMD, as previously reported in this sample 25 other samples.50-53
This was the first observation of a relationship between physical activity and early stage of AMD. Protective associations with physical activity were reported in relation to the incidence of diagnosed AMD or photographically evident advanced AMD in three previous studies.14-16 Evidence from the present study indicates that the association of physical acitivity is independent of diet. The degree to which associations of physical activity to AMD might have reflected better diets in people with higher levels of physical activity had not been previously assessed, except in one study with limited dietary data.15
Smoking has been the one risk factor most consistently associated with a higher risk for AMD.13 The association between lifelong smoking and AMD in the present study was only marginal and further attenuated after further adjusting for diet and physical activity. Obesity has been associated with a higher prevalence or incidence of AMD less consistently, 9 but weight loss has been associated with reduced AMD presence.10 Only extreme obesity was associated with AMD in the present sample; this was not significant after multivariate adjustment and was almost completely attenuated after adjusting for the potential explanatory variables of mHEI and physical activity. It is unclear how much of the associations of these risk factors in past studies may have been attributed to that fact that smoking and obesity are more common among people with poor diets and people who exercise less. Previous studies did not adjust for both diet and physical activity. Moreover, even when adjusted for, statistically, some residual confounding can be expected due to imperfect measurement and failure to capture these exposures over long periods in adult life.
In the present study, the three-fold lowering of odds for AMD among women with a combination of healthy, compared with unhealthy, lifestyles suggests that a combination of healthy lifestyle practices might be more important in reducing AMD risk than a focus on one. These changes, collectively, may contribute to lowering of oxidative stress, inflammation, blood pressure and improving blood lipids all of which are thought to be pathogenic mechanisms which promote AMD. It is well known that smoking increases oxidative stress54 and expected that stopping smoking lessens it. Physical activity can also up-regulate antioxidant protection enzyme systems, so that it reduces oxidative stress, despite the fact that bouts of physical activity can increase oxidative stress in the short-term (reviewed 55). Improvements in diet and physical activity alone or in conjunction with a reduction in obesity can lessen oxidative stress as well (reviewed56).
Healthy lifestyles may lower AMD risk by lowering systemic inflammation which is widely thought to contribute to AMD pathogenesis. Healthy diet patterns and physical activity have been related to lower blood levels of CRP, a marker of systemic inflammation in other samples, 31,57 as they were in the present sample (Table 3).
Healthy lifestyles may also lower AMD risk by reducing blood pressure(related to AMD risk in some past studies (previously reviewed 8). Intervention trials have demonstrated that reductions in blood pressure can result from healthy diets,58 physical activity,59 and weight loss.59, 60 A history of hypertension was less common in women in the highest quintiles for mHEI, physical activity and HLS-score (Table 3) .
In addition to these mechanisms, we speculate that healthy diets and physical activity might lower risk for AMD by improving status of macular pigment. Macular pigment density was associated with healthy diets, physical activity and HLS (Table 3) . The carotenoids which comprise macular pigment can block the frequencies of blue light that are known to damage the retina directly; they may also quench reactive oxygen species that form as a result of the light and oxygen rich environment (previously reviewed61) and they could reduce the formation of a toxic metabolite of retinal recycling (A2E) which is stimulated by blue light62 by blocking blue light from reaching the retina. Clearly, lutein and zeaxanthin supplementation from foods can increase macular pigment density, but the ability to increase macular pigment varies considerably among persons.63-65 As, we have previously discussed66 67 several aspects of diet, such as the overall intake of fruits, vegetables, whole grains, and fat may contribute to the uptake and turnover of these carotenoids. Physical activity might contribute to more dense macular pigment by reducing inflammation and oxidative stress, directly, or by reducing obesity. Obesity is related to lower macular pigment in this and other samples66, 68, 69 and may increase oxidative stress and carotenoid turnover, as well.56
Confirmation of these associations of healthy diets and lifestyles to AMD in intervention studies and long-term population-based studies which include men and a broader sample of ethnic backgrounds would provide additional evidence and more reliable risk estimates for the strong associations we observed in the present study among women. Our estimates in this primarily white sample may overestimate the overall impact in people from Hispanic, African and Asian origins who seem to be at lower risk for developing advanced stages of AMD, despite similar levels of early AMD.70 Conversely, estimates in the present study could be underestimated because women with less healthy lifestyles were less likely to participate in this study, weakening the power to estimate AMD rates among those with unhealthy lifestyle habits.
The HLS was not constructed a priori and needs to be further studied in separate samples. Healthy diets or lifestyles that we evaluated might reflect other unknown and unmeasured aspects of lifestyle. Socioeconomic status can be a surrogate for some such unknown factors. The socioeconomic status of women in this sample is high, limiting the extent to which this may be a confounder. For example, 78% of women had more than a high school education. Further adjustment for education or income levels did not influence associations in this sample (data not shown).
The size of this sample and of any single sample available today are not large enough, nor the sample diverse enough, to evaluate associations of each healthy lifestyle behavior to AMD independently. For example, we could not evaluate the potential benefit of healthy diets in women who were overweight compared with those who were lean. Exploratory analyses indicated stronger associations of mHEI-score to AMD among women with BMI>30, compared to <25, but these associations were not significant, nor were the interactions between diet and BMI significant. Larger studies or pooled samples across many studies might be useful to estimate interactions between these healthy habits and their influence on the occurrence of AMD.
Finally, it may be that the impact of healthy habits is more or less important in people who have high risk genotypes for AMD. In several recent studies healthy lifestyles have been more strongly associated with risk lowering among people with high variants of CFH Y402H 71 , 72-75 and ARMS2 A69S. 71 In the present study, having a family history for AMD did not modify the association of healthy diet or lifestyle with AMD, but genotyping will better characterize a person's susceptibility for the disease and improve the ability to examine the possibility that diet and lifestyle modifies genetic risk.
A combination of healthy lifestyle behaviors that includes healthy diets, physical activity and not smoking was associated with markedly lowered prevalence of early AMD, an average of six years later, in postmenopausal women. Adopting these healthy habits may markedly lower the prevalence of early stages of AMD, the number of people who develop advanced stages of AMD in their lifetime and health care costs associated with treatment for this condition.
These results also serve to remind us that risk for AMD is passed to subsequent generations not only through genes, but possibly also through the lifestyle habits we model and encourage. Specifically, we believe that these results, together with current scientific evidence for chronic disease prevention, support recommendations to exercise (“move” at least a low intensity for one to two hours per day; outside when possible), avoid smoking and follow a healthy diet pattern diets which is 1) abundant in plant foods (vegetables that include dark leafy green and orange, fruits and whole grains) 2) contains daily protein sources in moderation and variety (beans, nuts, fish, dairy, eggs , meat and poultry) and 3) limits foods which are high in sugar, fat, alcohol, refined starches and oils.
We thank the women who generously contributed their time to participate in the CAREDS. We also thank Rachel Adler for her work in computing diet scores.
Program Office: (National Heart, Lung, and Blood Institute, Bethesda, Maryland) Jacques Rossouw, Shari Ludlam, Joan McGowan, Leslie Ford, and Nancy Geller.
Clinical Coordinating Center: (Fred Hutchinson Cancer Research Center, Seattle, WA) Ross Prentice, Garnet Anderson, Andrea LaCroix, Charles L. Kooperberg; (Medical Research Labs, Highland Heights, KY) Evan Stein; (University of California at San Francisco, San Francisco, CA) Steven Cummings.
Clinical Centers: (Albert Einstein College of Medicine, Bronx, NY) Sylvia Wassertheil-Smoller; (Baylor College of Medicine, Houston, TX) Haleh Sangi-Haghpeykar; (Brigham and Women's Hospital, Harvard Medical School, Boston, MA) JoAnn E. Manson; (Brown University, Providence, RI) Charles B. Eaton; (Emory University, Atlanta, GA) Lawrence S. Phillips; (Fred Hutchinson Cancer Research Center, Seattle, WA) Shirley Beresford; (George Washington University Medical Center, Washington, DC) Lisa Martin; (Los Angeles Biomedical Research Institute at Harbor- UCLA Medical Center, Torrance, CA) Rowan Chlebowski; (Kaiser Permanente Center for Health Research, Portland, OR) Erin LeBlanc; (Kaiser Permanente Division of Research, Oakland, CA) Bette Caan; (Medical College of Wisconsin, Milwaukee, WI) Jane Morley Kotchen; (MedStar Research Institute/Howard University, Washington, DC) Barbara V. Howard; (Northwestern University, Chicago/Evanston, IL) Linda Van Horn; (Rush Medical Center, Chicago, IL) Henry Black; (Stanford Prevention Research Center, Stanford, CA) Marcia L. Stefanick; (State University of New York at Stony Brook, Stony Brook, NY) Dorothy Lane; (The Ohio State University, Columbus, OH) Rebecca Jackson; (University of Alabama at Birmingham, Birmingham, AL) Cora E. Lewis; (University of Arizona, Tucson/Phoenix, AZ) Cynthia A. Thomson; (University at Buffalo, Buffalo, NY) Jean Wactawski-Wende; (University of California at Davis, Sacramento, CA) John Robbins; (University of California at Irvine, CA) F. Allan Hubbell; (University of California at Los Angeles, Los Angeles, CA) Lauren Nathan; (University of California at San Diego, LaJolla/Chula Vista, CA) Robert D. Langer; (University of Cincinnati, Cincinnati, OH) Margery Gass; (University of Florida, Gainesville/Jacksonville, FL) Marian Limacher; (University of Hawaii, Honolulu, HI) J. David Curb; (University of Iowa, Iowa City/Davenport, IA) Robert Wallace; (University of Massachusetts/Fallon Clinic, Worcester, MA) Judith Ockene; (University of Medicine and Dentistry of New Jersey, Newark, NJ) Norman Lasser; (University of Miami, Miami, FL) Mary Jo O'sullivan; (University of Minnesota, Minneapolis, MN) Karen Margolis; (University of Nevada, Reno, NV) Robert Brunner; (University of North Carolina, Chapel Hill, NC) Gerardo Heiss; (University of Pittsburgh, Pittsburgh, PA) Lewis Kuller; (University of Tennessee Health Science Center, Memphis, TN) Karen C. Johnson; (University of Texas Health Science Center, San Antonio, TX) Robert Brzyski; (University of Wisconsin, Madison, WI) Gloria E. Sarto; (Wake Forest University School of Medicine, Winston-Salem, NC) Mara Vitolins; (Wayne State University School of Medicine/Hutzel Hospital, Detroit, MI) Michael S. Simon.
Women's Health Initiative Memory Study: (Wake Forest University School of Medicine, Winston-Salem, NC) Sally Shumaker.
SOURCES OF SUPPORT
This research was supported by The National Institutes of Health and The National Eye Institute Grants EY013018 and EY016886 and Research to Prevent Blindness. The WHI program is funded by the National Heart, Lung, and Blood Institute, National Institutes of Health, U.S. Department of Health and Human Services through contracts N01WH22110, 24152, 32100-2, 32105-6, 32108-9, 32111-13, 32115, 32118-32119, 32122, 42107-26, 42129-32, and 44221.