点击显示 收起
1 From the Jean Mayer US Department of Agriculture Human Nutrition Research Center on Aging at Tufts University, Boston, MA (C-JC and AT), and the AREDS Coordinating Center, The EMMES Corporation, Rockville, MD (RCM and GG)
2 Any opinions, findings, conclusions, or recommendations expressed in this publication are those of the authors and do not necessarily reflect the views or policies of the US Department of Agriculture, nor does mention of trade names, commercial products, or organizations imply endorsement by the US Government. 3 Supported by the US Department of Agriculture under agreements 1950-5100-060-01A (C-JC and AT) and R01-13250 and R03-EY014183-01A2 from the National Institutes of Health (AT) and by grants (to AT) from the Johnson and Johnson Focused Giving Program, D Gierhardt, and the Singers. The funding sources had no role in the design and conduct of the study; the collection, analysis, and interpretation of the data; or the preparation, review, or approval of the manuscript. 4 Address reprint requests to A Taylor, Jean Mayer USDA Human Nutrition Research Center on Aging at Tufts University, 711 Washington Street, Boston, MA 02111. E-mail: allen.taylor{at}tufts.edu.
ABSTRACT
Background:Age-related macular degeneration (AMD) is the major cause of irreversible blindness. AMD appears to share several carbohydrate-related mechanisms and risk factors with diabetes-related diseases, including retinopathy and cardiovascular disease (CVD); however, to date, only one small study has addressed this issue.
Objective:The objective was to test the hypothesis that dietary glycemic index (dGI), which has been related to the risk of diabetes and CVD, is associated with the risk and severity of AMD in nondiabetic elderly populations.
Design:Dietary information was obtained from 4099 participants aged 55–80 y (56% women) in the Age-Related Eye Disease Study (AREDS). A total of 8125 eligible eyes at baseline were classified into 1 of 5 AMD groups according to the size and extent of drusen, the presence of geographic atrophy, and neovascular changes. We used a generalized estimating approach to evaluate the relations between dGI and risk and severity of AMD with eyes as the unit of analysis.
Results:Compared with eyes in the first quintile of dGI, eyes in the fourth and fifth quintiles had a significantly or suggestively higher risk of large drusen, geographic atrophy, and neovascularization. The multivariate-adjusted odds ratios (95% CIs) for the highest quintile were 1.42 (1.09, 1.84), 1.78 (0.81, 3.90), and 1.41 (0.95, 2.08), respectively, of which only the odds ratio for large drusen was significant. A significant positive relation between dGI and severity of AMD was also noted (P for trend < 0.001). There was a 49% increase in the risk of advanced AMD (geographic atrophy plus neovascularization) for persons with a dGI higher than the sex median (women: 77.9; men: 79.3). This result indicated that 20% of prevalent cases of AMD would have been eliminated if the AREDS participants consumed diets with a dGI below the median.
Conclusion:The association between dGI and AMD from the AREDS cross-sectional analysis at baseline suggests that a reduction in the dGI, a modifiable risk factor, may provide a means of diminishing the risk of AMD.
Key Words: Retina • nutrition • carbohydrate • diabetes • insulin • cardiovascular diseases • glycation • fat • inflammation • aging • stress • epidemiology • risk factor • insulin-like growth factor
INTRODUCTION
Age-related macular degeneration (AMD) is a multifactorial, neurodegenerative disease of the central retina or macula. The macula comprises only 4% of the total human retina area, but it is responsible for all our high-acuity vision. AMD usually occurs after middle age and is the leading cause of irreversible vision loss in Australian, Western European, and North American populations (1). As populations in these developed countries continue to age, this condition is emerging as a major public health issue. In the United States alone, the number of people with visually impairing AMD is expected to double and reach 3 millions by 2020 (2). There is no effective therapy for AMD, but dietary intervention (eg, antioxidant supplementation) appears to offer a means to delay the progression of this debility (3, 4). In contrast with the many studies of antioxidant nutrient intakes that have been conducted (5, 6), only one epidemiologic study of the associations between dietary carbohydrate and risk of AMD has been conducted (7). This information is particularly important because there has been an increasing trend of carbohydrate intake in the United States during the past 30 y, and one-half of our calories come from carbohydrates (8). Furthermore, increasing evidence indicates that carbohydrate can damage ocular tissues (9) and that dietary carbohydrate is associated with the risk of age-related eye diseases in nondiabetic persons (7, 10, 11). This is not surprising for the following reasons: 1) despite differences in pathological features between age-related and diabetes-related eye diseases, they share several carbohydrate-related mechanisms (9), including the formation of advanced glycation end products (AGE) and their sequelae (7, 9, 12, 13); 2) hyperglycemia-mediated damage can occur below the diabetic threshold in diabetes and cardiovascular disease (CVD), for example (14, 15), and AMD and CVD appear to share some risk factors (16); and 3) ocular tissues are totally dependent on the circulation for the glucose supply. Because carbohydrate nutrition has been linked to the development of diabetes and CVD, we speculate that dietary carbohydrate may also play a role in the development of AMD.
The glycemic index (GI) is a physiologic measure of the glycemic quality of carbohydrate-containing foods and can be used to guide consumers to choose foods. High-GI foods result in elevated blood glucose concentrations relative to low-GI foods. GI is defined as the ratio of area under curve of 2-h blood glucose curves from the same amount (50 g) of available carbohydrate from test food versus reference food (pure glucose or white bread) (17). The dietary GI (dGI) is a weighted average of the GIs of foods in the diet [ (GIi x Wi)/W] (18). dGI has been implicated in the development of obesity, diabetes, CVD, and cancers (19, 20). Studies also suggest that a high dGI is associated with several components of the metabolic syndrome, such as low HDL cholesterol (21, 22), and with measures of chronic inflammation, such as elevated C-reactive protein concentrations (23), which have been related to CVD and to AMD (24-27).
Compared with the previous study, which addressed only the earliest manifestations of AMD (7), in the present study we studied the association of dGI with early and late stages of AMD in nondiabetic elderly participants in the Age-Related Eye Disease Study (AREDS; n = 4757 persons aged 55–80; 56% women) (28).
SUBJECTS AND METHODS
AREDS population
AREDS of the National Eye Institute of the National Institutes of Health (Bethesda, MD) is a long-term multicenter, prospective study dedicated to assess the clinical course, prognosis, risk factors, and prevention strategy of both AMD and cataract (28). The protocol was approved by a Data and Safety Monitoring Committee and by each Institutional Review Board for the 11 participating ophthalmic centers before initiation of the study. Participants were 55–80 y of age at enrollment and were required to have at least one eye with a visual acuity of 20/32 or better, and the lens and vitreous had to be sufficiently clear to permit good quality retinal photographs that would permit identification and quantification of small drusen. In addition, at least one eye of each participant was free from eye disease that could complicate assessment of AMD or lens opacity progression (eg, optic atrophy and acute uveitis), and that eye could not have had previous ocular surgery (except cataract surgery and unilateral photocoagulation for AMD). Finally, potential participants were excluded for illness or disorders that would have made long-term follow-up or compliance with the study protocol unlikely or difficult. A total of 4757 participants were enrolled from November 1992 to January 1998. Informed consent was obtained from participants before enrollment.
Procedures
Data on possible risk factors for AMD were obtained from a baseline general physical and ophthalmic examination, a detailed questionnaire on basic characteristics and demographic data, and a food-frequency questionnaire (FFQ) (29). The FFQ was validated in relation to a 24-h dietary recall in a subset of the AREDS volunteers (n = 192). Correlations for energy and carbohydrate intakes between the 24-h dietary recall and the FFQ were 0.51 (P < 0.001) and 0.56 (P < 0.001), respectively (N Kurinij, G Gensler, and R Milton, unpublished observation, 1998). Stereoscopic fundus photographs of the macula and slit lamp and red reflex lens photographs were taken and graded at a central ophthalmic photograph reading center, where the various lesions associated with AMD and the degree of lens opacities by type were assessed according to AREDS grading procedures adapted from the Wisconsin age-related maculopathy grading system and the Wisconsin System for Classifying Cataracts from Photographs, respectively (30, 31). The AREDS AMD Classification System demonstrated satisfactory reliability for detecting onset of advanced AMD and moderate to substantial agreement on various abnormalities across the AMD spectrum (32).
For AMD grading, eyes were classified into 1 of 5 groups (see below) according to the size and extent of drusen, the presence of geographic atrophy, and neovascular changes of AMD (32). The 5 groups, numbered serially and based on increasing severity of drusen or type of AMD, were defined as follows: group 1 (control), eyes had no drusen or nonextensive small drusen (n = 2750 eyes); group 2 (intermediate drusen), eyes had one or more intermediate drusen, extensive small drusen, or pigment abnormalities associated with AMD (n = 1806 eyes); group 3 (large drusen), eyes had one or more large drusen or extensive intermediate drusen (n = 2803 eyes); group 4 (geographic atrophy), eyes had geographic atrophy (n = 164 eyes); and group 5 (neovascular), eyes had choroidal neovascularization or retinal pigment epithelium detachment (n = 602 eyes).
Study subjects
The recruitment scheme of AREDS used in the present study is illustrated in Figure 1. Of the available 4757 subjects, we first excluded 658 persons, including 398 persons with diabetes, 99 persons with invalid calorie intakes (valid intakes ranged from 400 to 3000 kcal/d for the women and from 600 to 3500 kcal/d for the men), and 161 persons with missing nutritional, nonnutritional, and ophthalmologic covariates. This left 8125 eyes from 4099 persons; 73 persons contributed only 1 eye because 35 left eyes and 38 right eyes had "disqualifying ocular lesions" or "could not be graded."
FIGURE 1.. Flow chart describing the disposition of subjects from the Age-Related Eye Disease Study (AREDS).
Assessment of dietary carbohydrate variables
A 90-item modified Block FFQ was administered to AREDS participants at baseline. The FFQ collected information about usual dietary intakes over the previous year and classified them into 9 possible response categories, ranging from "never or less than once per month" to "2 or more times per day." The daily total carbohydrate intake of an individual was calculated by summing the product of the frequency, serving size, and carbohydrate content per serving from individual food items derived from the nutrition database of the Nutrition Coordinating Center at the University of Minnesota. The GI (17) values for foods in the FFQ were either derived from published values based on white bread as the reference food or were imputed from GI values of comparable foods (33). The dGI for each subject was calculated as the weighted average of the GI scores for each food item, with the amount of carbohydrate consumed from each food item as the weight (18):
RESULTS
The distribution of baseline AREDS characteristics is shown in Table 1. The multivariate-adjusted associations (ORs and 95% CIs) between baseline characteristics and prevalence of AREDS AMD groups are shown in Table 2. Compared with group 1 (n = 2750 eyes), cases in group 2 (n = 1806) were significantly older, were less educated, and had more sunlight exposure. Cases in group 3 (n = 2803) were significantly older, less educated, more likely to be white, more likely to be a smoker, and more likely to have a hypertension history and lens opacity. Cases in group 4 (n = 164) were significantly older, less educated, and more likely to be a smoker. As noted previously (29), cases in group 5 (n = 602) were significantly older, less educated, more likely to be white, more likely to be a smoker, more likely to have a hypertension history, more likely to have a higher BMI, and more likely to have lens opacity and hyperopia.
View this table:
Table 1. Baseline characteristics of eyes by age-related macular degeneration (AMD) group (n = 8125)1
View this table:
Table 2. Associations [odds ratios (ORs)] between baseline characteristics and prevalence of age-related macular degeneration by macular degeneration group
Although all correlation coefficients (parenthesized) between dGI and carbohydrate (–0.04), fat (0.29), vitamin C (–0.23), vitamin E (0.15), zinc (–0.043), ß-carotene (–0.14), lutein and zeaxanthin (–0.08), folic acid (0.05), riboflavin (–0.24), niacin (0.22), thiamine (0.10), and energy (0.06) intakes were significant (P < 0.05), they were weakly correlated.
In the analysis of the association between dGI and AMD risk for each category of AMD, we found a generally similar pattern for AREDS groups 3, 4, and 5 (Figure 2). Compared with eyes in the first quintile of dGI, eyes in the fourth and fifth quintiles had significantly or suggestively higher risk. The multivariate-adjusted ORs (95% CIs) for the fourth and fifth quintiles, respectively, were 1.31 (1.02, 1.66) and 1.42 (1.09, 1.84) for AREDS group 3, 1.58 (0.79, 3.16) and 1.78 (0.81, 3.90) for AREDS group 4, and 1.28 (0.89, 1.84) and 1.41 (0.95, 2.08) (P = 0.0851) for AREDS group 5. The P value for the test of trend was 0.001 for group 3, 0.082 for group 4, and 0.005 for group 5. We did not find a significantly increased risk in AREDS group 2 with increased dGI.
FIGURE 2.. Associations [odds ratios (ORs) and 95% CIs] between dietary glycemic index (dGI) and prevalence of age-related macular degeneration by macular degeneration group. Participants were divided into quintile categories according to their dGI; those in the lowest 20% of the distribution comprised the referent category. The cutoffs were 73.6, 76.6, 79.1, and 81.7 for the women and 75.7, 78.3, 80.3, and 82.8 for the men. The multivariate-adjusted logistic models using group 1 (n = 2750) as a control were adjusted for age, sex, race, education, smoking status, BMI, sunlight exposure, hypertension, lens opacity, refractive error, and energy-adjusted dietary fat, lutein and zeaxanthin, folic acid, niacin, riboflavin, thiamine, ß-carotene, vitamin C, vitamin E, zinc, and total carbohydrate intakes.
In the analyses of interactions with sex, education, and smoking in group 3 or group 5, we found no significant interaction in either additive OR models or in stratification analyses. Using stratification analysis, we found no significant interaction with age in group 3 and 5. However, in the additive OR models, we found a marginally significantly (P = 0.099) synergistic response between the oldest age category (71 y) and the fifth quintile of dGI in group 3, but no significant interaction was found in group 5. In the interaction analysis with BMI using the additive OR model, we found that for BMIs >31 (P = 0.086) and 31–23.6 (P = 0.023), there was a synergistic response with the fifth dGI quintile on the risk for group 3. However, there was no significant interaction in the stratification analysis in group 3. In the interaction analysis in group 5, although we found no significant effect modification of BMI in the additive OR model, a significant interaction with BMI in the stratification analysis was found (P = 0.04); we found that the higher the BMI stratum, the higher the OR of neovascularization for high dGI (sex median): 2.45 (P = 0.001), 1.48 (P = 0.012), and 1.03 (P = 0.91) for BMI 31, 31–23.6, and <23.6, respectively.
In the multivariate analysis for total carbohydrate intake, no significant trend was found (Figure 3). The multivariate-adjusted mean and 95% CI of dGI for each AREDS AMD group is shown in Figure 4. The data show that the higher the severity of AMD, the higher the dGI. The multivariate-adjusted means (95% CIs) of dGI were 77.99 (77.87, 78.11) for AREDS group 1, 78.11 (77.96, 78.26) for group 2, 78.36 (78.24, 78.48) for group 3, 79.40 (78.90, 79.90) for group 4, and 79.01 (78.76, 79.26) for group 5 (Figure 4). Both of the P values for trend (group 1 group 2 group 3 group 4 and group 1 group 2 group 3 group 5) were <0.001.
FIGURE 3.. Associations [odds ratios (ORs) and 95% CIs] between total carbohydrate intake and prevalence of age-related macular degeneration by macular degeneration group. Participants were divided into quintile categories according to their total carbohydrate intake; those in the lowest 20% of the distribution comprised the referent category. The cutoffs were 130.1, 147.2, 162.7, and 182.4 g/d for the women and 151.2, 170.6, 187.9, and 209.0 g/d for the men. The multivariate-adjusted logistic models using group 1 (n = 2750) as a control were adjusted for age, sex, race, education, smoking status, BMI, sunlight exposure, hypertension, lens opacity, refractive error, dietary glycemic index, and energy-adjusted dietary fat, lutein and zeaxanthin, folic acid, niacin, riboflavin, thiamine, ß-carotene, vitamin C, vitamin E, and zinc intakes.
FIGURE 4.. Multivariate-adjusted mean dietary glycemic index (dGI) by age-related macular degeneration (AMD) groups. Error bars represent 95% CIs. Multiple linear regression was used and adjusted for age, race, sex, education, smoking status, BMI, sunlight exposure, hypertension, lens opacity, refractive error, AMD group, and energy-adjusted dietary fat, lutein and zeaxanthin, folic acid, niacin, riboflavin, thiamine, ß-carotene, vitamin C, vitamin E, zinc, and total carbohydrate intakes. P for trend values were <0.001 for controlintermediate drusenlarge drusengeographic atrophy and for controlintermediate drusenlarge drusenneovascular.
The OR of advanced AMD (group 4 plus group 5) for those with dGI versus < the sex median was 1.49 (95% CI: 1.19, 1.85). Using this estimate in Equation 2, we computed a prevalent PAF of 20% for advanced AMD for high dGI in the AREDS cohort at baseline.
DISCUSSION
The present study supports and strengthens our hypothesis that diets that provide a higher dGI are positively associated with the risk of AMD (7). We estimate that 20% (PAF) of prevalent cases of advanced AMD (group 4 plus group 5) in the AREDS cohort would be eliminated if the AREDS participants consumed diets that have dGI values below the median (38).
Previous study
Only one prior study, the cross-sectional analysis of the Nutrition and Vision Project (NVP) of the Nurses’ Health Study, related dietary carbohydrate to AMD (7). Both studies were mainly composed of whites and used the same classification system of AMD (7, 32). In comparison with participants in the NVP (mean age: 61 y), the present study had older participants (mean age: 68 y) and was much larger and thus offered a far greater number of cases and a more complete spectrum of the lesions that define AMD. In addition, the AREDS cohort allowed us to further evaluate the interactions with other risk factors. Although the NVP collected previous long-term dietary data, the AREDS evaluated the dietary data for the previous year. However, the present findings are consistent with the findings in the NVP (7). Drusen and pigmentary abnormality are considered early signs of AMD and strong predictors of subsequent advanced AMD (39). In the NVP study, a significantly positive association was noted between dGI and pigmentary abnormality, but we could not associate intermediate drusen with dGI, seemingly because of a lack of power (7). In the present study, pigmentary abnormality was classified into AREDS group 2. This precluded evaluating the specific associations with pigmentary abnormality. However, we had much more statistical power to evaluate the associations of interest than was possible in the NVP study, especially for drusen and advanced AMD. In the present study, we found positive associations between dGI and large drusen (group3) and advanced AMD (group 4 plus group 5). Given the results that indicate a significantly positive relation between dGI and severity of AMD, we speculate that dGI plays a role in both the early and late stages of AMD.
In the NVP, a one-unit difference in dGI was observed between early AMD (pigmentary abnormality) cases and controls, based on the means of the previous 10-y diet (7). Our present data indicate that advanced AMD cases have a mean dGI that is up to 1.4-units higher than that of controls, based on the diet over the previous year (Figure 4). According to Equation 1, the difference in dGI (dGI) can be calculated from [(GIi x Wi)]/W. For a person consuming 3 servings of bread (14 g available carbohydrate per serving) in a diet containing 250 g total available carbohydrate per day, a 1.4-unit reduction in dGI is achievable by replacing the 3 servings of white bread [(GI = 100, food item no. 101 (34)]) with 3 servings of whole-grain bread [GI = 89, food item no. 123 (34)] from the daily diet, where dGI = [(100 x 14 x 3) –(89 x 14 x 3)]/250 = 1.85. However, the difference in dGI between cases and controls may be an underestimate for the diet change associated with the reduced risk of AMD, likely because of survival bias. This is because dGI has been positively related to several major causes of mortality (19, 20). Prospective studies are needed to refine these data to arrive at dietary recommendations.
Possible mechanisms
Much evidence supports our hypothesis. First, the blood retina barrier expresses high concentrations of glucose transporters to satisfy the large demand for glucose metabolism and this facilitates the formation and accumulation of AGE (9, 40). AGE are thought to have deleterious effects on the activity of degradative enzymes, on retinal pigment epithelium function, and on the integrity of the choriocapillaris and Bruch's membrane (9). These biochemical and physiologic compromises could be especially pronounced after the consumption of high-GI foods, which induce an abrupt increase in blood glucose concentrations and thus may supply an excess amount of glucose relative to demand during the postprandial stage.
Second, all forms of drusen appear to have similar carbohydrate components and AGE accumulate in drusen with age and occur at a higher level in patients with advanced AMD (41). Third, a high-GI diet has been proposed to play a role in producing oxidative stress and exacerbating proinflammatory processes (19, 23, 42, 43). Oxidative insults may decrease the efficacy of the quality-control machinery in the retina (44). In addition, other hyperglycemia-mediated damage, including inflammatory and angiogenic responses that occur in advanced AMD, was found to at least partially explain AMD pathogenesis (45). It is also likely that the compensatory hyperlipidemia that follows hypoglycemia in the late postprandial stage after the consumption high-GI foods, and which is thought to be important in the development of CVD (19), may also play some role in AMD pathogenesis because hyperlipidemia has been implicated in the pathogenesis of AMD (26). Finally, the insulin-like growth factor axis, which plays an essential role in cell proliferation and differentiation and complements the metabolic effects of insulin, has been linked to aging and age-related diseases, including diabetes, cancers, CVD, retinopathy, as well as AMD (20, 46-48). Interestingly, recent evidence suggests that dGI could, through modulating the insulin-like growth factor axis, affect the risk of age-related diseases (20, 46). Further studies are needed to clarify the mechanism.
In the interaction analysis, we found a potential synergistic effect between old age and high dGI on the risk for group 3 (large drusen/extensive intermediate drusen). This finding implies that dGI may partially, in addition to the biological compromise with age itself, explain the observation that drusen accumulates with age. A synergistic response was also noted between high BMI and high dGI in Group 3. In the stratification analysis of group 5, we found that the ORs for high dGI were larger in the high-BMI strata than in the low-BMI strata, which implies that dGI has a relatively higher effect on the risk of neovasularization in persons with a high BMI. However, because of inherent limitations in studying interaction with the use of epidemiologic data, studies adopting different designs and settings are needed to clarify the detailed mechanisms and implications (37).
Strengths and limitations
Using participants from a well-characterized cohort, we were able to use the standardized collection of risk factor information and photographic grading of maculopathy. Using eyes as the unit of analysis, we increased our power, which improved our ability to adjust for many previously identified risk factors and potential dietary confounders and to evaluate the interactions of interest in our multivariate analysis. This approach was further supported by the observation that, despite the difference in the unit of analysis [by person in the previous AREDS report (29) and by eye in the present study], the risk factor profiles were very similar in this study and in the previous AREDS report. Old age, lower education, and smoking were the 3 most important risk factors for AMD in the AREDS cohort. Recall and selection bias in the AREDS were unlikely to explain our findings because exposure information was collected before outcome evaluation, and our retinal classifications were performed in an independent center by graders masked to our nutrition data (11). Although GI values are generally reproducible from place to place, there are some variations in published GI values for apparently similar foods (33). For these foods, we chose the GI of the most popular American food item in our compilation (11). It is unlikely that the nondifferential misclassification in our dGI compilation could explain our findings because our compilers were blinded to the ophthalmic data. Consistency with prior evidence reduced the possibility that the present findings were due to chance. Residual confounding is a concern but should be minimized because we included all known dietary and nondietary confounders in our analysis. The low correlations between dGI and other nutritional covariates indicated that it is unlikely that any other single nutrient or dietary pattern could totally account for the independent effect of dGI on the risk of AMD. The cross-sectional nature of this study limited its strength in defining causality and dietary recommendation.
Conclusions
In summary, these cross-sectional analyses suggest that poor dietary carbohydrate quality as defined by dGI, a modifiable risk factor, may increase the risk of AMD through several common etiologic factors of diabetes and CVD, including the formation of AGE and increases in oxidative stress, inflammation, and hyperlipidemia. Our results also suggest that the quality, but not the quantity, of dietary carbohydrate influences the risk of AMD in both the early and late stages of the disease. Prospective studies are needed before dietary carbohydrate management is recommended as another strategy for the prevention of AMD.
ACKNOWLEDGMENTS
The authors’ responsibilities were as follows—C-JC and GG: had full access to all of the data and took responsibility for the integrity of the data and the accuracy of the data analysis; C-JC, RCM, and AT: conceived and designed the study, analyzed and interpreted the data, and critically analyzed the manuscript for important intellectual content; RCM: acquired the data; C-JC and AT: drafted the manuscript; CJC: conducted the statistical analysis; GG: provided administrative, technical, and material support; RCM and AT: supervised the study; all authors: contributed substantially to the manuscript. No conflicts of interest were declared.
REFERENCES