Chapter 63 Epidemiology and Risk Factors for Age-Related Macular Degeneration
Age-related macular degeneration (AMD) is the leading cause of irreversible blindness.1 The disease adversely affects quality of life and activities of daily living, causing many affected individuals to lose their independence in their retirement years. AMD is estimated to affect more than 8 million individuals in the USA;2 the advanced form of the disease affects more than 1.75 million individuals.1 Despite the introduction of new therapies for prevention and treatment of AMD, the prevalence of AMD is expected to increase by 97% by the year 2050.3
The only proven treatment available for the dry or nonexudative forms of this disease, comprising 85% of cases, is an antioxidant/mineral supplement which can slow the progression of the disease by 25% over 5 years.4 For the wet form of the disease, anti-vascular endothelial growth factor (VEGF) treatments have been very effective in preventing severe vision loss. Still, preventive measures are needed to reduce the burden of this disease. Smoking is the most consistently identified modifiable risk factor.5–6 Obesity, sunlight exposure, and nutritional factors including antioxidants and dietary fat intake may also affect AMD incidence and progression.4,7–14 There has also been great progress in identifying the genetic variants that impact risk of AMD.15–29 The knowledge of genetic risk variants for the disease coupled with knowledge of nongenetic risk factors have improved the ability to predict which patients will develop advanced forms of the disease.23,30–32 Although much progress has been made over the past two decades, finding the causes and mechanisms of this condition remains a challenge.
Macular degenerative changes have typically been classified into two clinical forms, dry or wet, both of which can lead to visual loss. The wet form is also called advanced wet, exudative or neovascular. In the early or intermediate dry forms visual loss is infrequent, and when it occurs it is usually gradual. Ophthalmoscopy reveals yellow subretinal deposits called drusen, or retinal pigment epithelial (RPE) irregularities, including hyperpigmentation or hypopigmentary changes. Larger drusen may become confluent and evolve into drusenoid RPE detachments. These drusenoid RPE detachments often progress to geographic atrophy and less frequently to neovascular AMD. Geographic atrophy involving the center of the macula, which is the advanced dry form, leads to visual loss. Each of these signs can be further subdivided according to the number or size of the lesions. In the wet form, vision loss can appear to occur suddenly, when a choroidal neovascular membrane leaks fluid or blood into the subpigment epithelial or subretinal space. Serous RPE detachments with or without coexisting choroidal neovascularization (CNV) are also classified as the wet form. Exudative serous RPE detachments often, but not always, advance to the neovascular stage. This phenotypic heterogeneity, or wide range of clinical findings, has led to the use of various definitions of AMD and also to some difficulties with comparisons among studies.
It is important for investigators to standardize definitions of a disease and its subtypes to enhance comparability and to promote collaborative efforts. Toward this goal, an international classification and grading system for AMD was recommended, although it is not universally applied.33 In this system age-related maculopathy (ARM) or early AMD is defined as the presence of drusen and RPE irregularities, and the terms late ARM and advanced AMD are limited to the occurrence of geographic atrophy and neovascular disease, the forms most often associated with greater visual loss. Clinical manifestations of AMD can be subcategorized according to the specific type of AMD, which for example can yield a four- or five-step grading system.34–35 The Clinical Age-Related Maculopathy grading System (CARMS)34 is useful for clinical management and genetic epidemiologic research and has been used in several studies.11,13,23,25–27,29,36–39 Alternative and more detailed systems have been used in some of the population-based studies described below.40–41 New subcategories of AMD will evolve as genetic and epidemiologic studies provide further insight into the pathogenesis of this disease. An updated classification under the auspices of the Beckman Initiative for Macular Research is underway.
Prevalence is the total number of cases in the population, divided by the number of individuals in the population. Population-based studies that have provided information on the prevalence of AMD within the USA include the National Health and Nutrition Examination Survey (NHANES),42–43 the Framingham Eye Study (FES),44 the Chesapeake Bay Watermen Study,40 the Beaver Dam Eye Study (BDES),45 the Baltimore Eye Survey,46 and the Salisbury Eye Evaluation Project.47 Population-based studies outside the USA include the Rotterdam Study in the Netherlands,48 the Blue Mountains Eye Study (BMES) in Australia,49 the Barbados Eye Study,50 and a study in Italy.51 Prevalence rates are quite variable for all types of AMD combined, because of differences in definitions of AMD, but are more consistent for “advanced AMD.”
The BDES was a census of the population of Beaver Dam, Wisconsin.45 This study found that the early forms are much more common than the late stages of AMD, and both types increase in frequency with increasing age. The prevalence of late AMD was 1.6% overall; exudative maculopathy was present in at least one eye in 1.2% of the population; and geographic atrophy was present in 0.6%. The prevalence of late AMD rose to 7.1% in persons who were 75 or older.
Total prevalence of AMD in the USA was also estimated in 2004 using pooled findings from seven large population-based studies both inside and outside the USA, and applying those prevalence rates to the USA population.1 This meta-analysis by the Eye Diseases Prevalence Group calculated the overall prevalence of neovascular AMD and/or geographic atrophy to be 1.47% of the USA population aged 40 years or older. The most recent NHANES, conducted from 2005 to 2008, sampled approximately 5500 persons.43 The total prevalence of any AMD in this civilian noninstitutionalized USA population aged 40 years or older was 6.5% (7.2 million people), and 809 000 persons were estimated to have the late stage of AMD.43
Studies conducted outside the USA have found similar or lower rates of AMD compared to those conducted inside the USA. In the Rotterdam Study, fundus photographs of 6251 participants aged 55–98 years were reviewed for drusen, pigmentary changes, and atrophic or neovascular AMD.48 The prevalence of AMD was observed to be slightly lower in that study compared with the BDES in Wisconsin. In the BMES in Australia, the authors also found lower prevalence of all lesions related to AMD in each age stratum.49 After adjusting for age, differences were significant for both soft drusen and retinal pigmentary abnormalities; they were lower but not significantly different for geographic atrophy and exudative disease. In a population-based study of 354 participants in rural southern Italy, the prevalence rates of AMD were also lower than those found in the USA.51 Methodological differences between studies may exist, but the lower prevalence rates found in these countries may also reflect genetic or environmental differences compared with the US population.
Incidence is a measure of the risk of developing some new condition within a specified period of time. A few studies have been done to evaluate the incidence of AMD. The FES used the age-specific prevalence data to estimate 5-year incidence rates of AMD, according to the definition of AMD in that study. These estimates were 2.5%, 6.7%, and 10.8% for individuals who were 65, 70, and 75 years of age, respectively.52 The BDES determined the 5-year cumulative incidence of developing early and late AMD in a population of 3583 adults (age range 43–86 years).53 Incidence of early AMD increased from 3.9% in individuals aged 43–54 years to 22.8% in persons 75 years of age and older. The overall 5-year incidence of late AMD was 0.9%. Persons 75 years of age or older had a 5.4% incidence rate of late AMD. The Visual Impairment Project of Melbourne, Australia, described the 5-year incidence of early AMD lesions in a population of 3271 participants aged 40 years and older.54 The overall 5-year incidence of AMD was 0.49%, and overall incidence of early AMD was 17.3% in this population. As with the BDES, incidence of AMD increased with age – up to 6.3% for people aged 80 years and older at baseline. The Barbados Eye Study described a 4-year incidence of early macular changes as 5.2% in a black population, with an extremely low incidence of exudative AMD.55 The differences in prevalence and incidence rates by race/ethnicity are discussed below. One report suggests that the incidence of advanced AMD in the USA may be on the decline, possibly due in part to changes in lifestyle habits of the American public over the past 40 years.56
The psychologic costs associated with AMD underscore the growing importance of this disease on the expanding older adult population. For this reason it is important to incorporate a functional component into studies of AMD. Instruments such as the National Eye Institute Visual Function Questionnaire and the Macular Disease Dependent Quality of Life Questionnaire have been used in AMD studies.57–58 Patients with visual loss resulting from AMD often report AMD as their worst medical problem and have a diminished quality of life.59–60 More recently in one study of well-being, patients with AMD had lower scores than patients with chronic obstructive pulmonary disease and acquired immunodeficiency syndrome (AIDS); the lower quality of life in patients with AMD was related to greater emotional distress, worse self-reported general health, and greater difficulty carrying out daily activities. Not only is AMD associated with a higher rate of depression in the community-dwelling adult population when compared to the unaffected adult population,61–62 but depression also exacerbates the effects of AMD.63
All studies demonstrate that the prevalence, incidence, and progression of all forms of AMD rise steeply with increasing age. There was a 17-fold increased risk of AMD comparing the oldest to the youngest age group in the Framingham Study.44 In the Watermen Study, the prevalence of moderate to advanced AMD doubled with each decade after age 60.40 In the BDES, approximately 30% of individuals 75 years of age or older had early AMD; of the remainder, 23% developed early AMD within 5 years.45,53By age 75 years and older in that study, 7.1% had late AMD, compared with 0.1% in the age group 43–54 years and 0.6% among persons aged 55–64 years. Pooled data in a prevalence paper showed similar rates, with dramatic increases in rates for both men and women older than 80 years.1
Several studies1,44–45,48 have shown no overall difference in the frequency of AMD between men and women, after controlling for age. However, in NHANES III, men, regardless of race and age, had a lower prevalence of AMD than women.42 Incidence rates within the Beaver Dam population also suggest a gender difference. After adjusting for age, women aged 75 years or older had approximately twice the incidence of early AMD compared with men.53 A study using reported incidence of exudative AMD in the USA among Medicare beneficiaries supported the Beaver Dam results.64 In the BMES, there were consistent, though not significant, gender differences in prevalence for most lesions of AMD, with women having higher rates for soft, indistinct drusen, but not for retinal pigmentary abnormalities.49 A case–control study in the Age-Related Eye Disease Study (AREDS) also found women had a higher risk for intermediate drusen.65 Residual confounding by age in the broad age category “75 and older” may partially explain the differences between studies since there are more women than men in that age group. Additional research is needed, however, to assess these associations.
Ophthalmologists observe visual loss caused by CNV less frequently among US ethnic minority groups compared with Caucasians. In the Baltimore Eye Survey, AMD accounted for 30% of bilateral blindness among whites and for 0% among African Americans.66 Data from a population-based study of blacks in Barbados, West Indies,50,55 revealed that incidence of AMD and signs of AMD changes occurred commonly but at a lower frequency than in predominantly white populations in other studies. Hispanics also have a lower prevalence of advanced AMD than non-Hispanics. Late-stage AMD was significantly less frequent among Hispanics vs. non-Hispanic whites in Beaver Dam (OR = 0.07; 95% CI = 0.01–0.49).67 The Los Angeles Latino Eye Study indicates Latinos have a relatively high rate of early AMD but not late AMD.68 Among persons aged 40–79 years, the age-specific prevalence of late AMD in Asians was comparable with that reported from white populations, but early AMD signs were less common among Asians.69
Overall, the literature to date suggests that early AMD is common among blacks and Hispanics, although less common than among non-Hispanic whites, whereas advanced AMD is much less common in these groups compared with non-Hispanic whites. Furthermore, differences in prevalence rates between non-Hispanic whites in different regions of the USA suggest that ethnicity is an important determinant of AMD.
Less education and lower income have been shown to be related to increased morbidity and mortality from a number of diseases,70 and there are mixed findings for AMD. The Eye Disease Case Control Study (EDCCS), a National Eye Institute-sponsored multicenter study, was designed to study risk factors for several types of maculopathy, including neovascular AMD.71 Persons with higher levels of education had a slightly reduced risk of neovascular AMD, but the association did not remain statistically significant after multivariate modeling.71 Education was also inversely related to AMD in case–control and prospective studies based on the AREDS population even in multivariate analyses.30,65 In the BDES, no association was found between education, income, employment status, or marital status and maculopathy.72 Furthermore, no associations were noted in another case–control study73 or in the FES,44 although different definitions of macular degeneration were used in those reports, compared with the more recent studies. It is possible that education is a surrogate marker for behaviors and lifestyles related to AMD.
Several case–control studies have shown an association between AMD and hyperopia.65,71,73–75 The potential problem with some of these studies is the clinical setting in which they were conducted. Because ophthalmology practices tend to contain a disproportionate number of myopic patients, controls selected from such practices would tend to have a higher prevalence of myopia than that of the general population. Population-based data from the BMES, less likely to have such potential bias, has also suggested a weak association between hyperopia and early AMD, but not late AMD.76 The population-based Rotterdam Study also showed an association between hyperopia and both incident and prevalent AMD.77 This association, therefore, might implicate structural and mechanical differences that render some eyes predisposed to maculopathy.78
Higher levels of ocular melanin may be protective against light-induced oxidative damage to the retina, since melanin can act as a free radical scavenger and may have an antiangiogenesis function. To date, the literature is inconclusive about the relationship between iris color and AMD. Darker irides have been found to be protective in some studies73,79–83 but not in others.71,84–88 Differences between studies may be partly related to the use of different definitions of disease, different number and types of other factors evaluated simultaneously, and residual confounding by ethnicity in some studies.
Data regarding the relationship between cataracts and AMD are inconsistent. FES investigators found no relationship,89 whereas data from the NHANES did support a relationship between AMD and lens opacities.90 In the BDES, in which photographs of the lens and macula were graded, nuclear sclerosis was associated with increased odds of early AMD (OR 1.96; 95% CI 1.3–3.0) but not of late AMD. Neither cortical nor posterior subcapsular cataracts were related to AMD.91 A case–control study of 1844 cases and 1844 controls indicated that lens opacities or cataract surgery were associated with an increased risk of AMD.74
Although AMD-affected individuals reported better visual function and quality of life after cataract surgery,92 a history of cataract surgery has been found to be associated with an increased risk for advanced AMD in some earlier studies.93 Investigators have postulated that this association might arise because the cataractous lens can block damaging ultraviolet light. Inflammatory changes after cataract surgery may also cause progression of early to late AMD. In the BDES, previous cataract surgery at baseline was associated with a statistically significant increased risk for progression of AMD (OR 2.7) and for development of late AMD (OR 2.8; 95% CI 1.03–7.6).85 In more recent prospective studies, however, including the large AREDS study cohort, there was no evidence to support a higher rate of progression of AMD in patients who underwent cataract surgery.94–95
The EDCCS demonstrated that eyes with larger cup-to-disc ratios had a reduced risk of exudative AMD. This effect persisted even after multivariate modeling,71 adjusting for known and potential confounding factors. Whether this finding, which is consistent with the association between AMD and hyperopic refractive error mentioned earlier, is meaningful in terms of the mechanisms associated with the development of AMD awaits further study.
The preponderance of epidemiologic evidence indicates a strong positive association between both wet and dry AMD and smoking. Two large prospective cohort studies have evaluated the relationship between smoking and wet AMD and dry AMD associated with visual loss.5,96 Seddon et al. reported that women in the Nurses’ Health Study who currently smoked 25 or more cigarettes per day had a relative risk (RR) of 2.4 (95% CI 1.4–4), and women who were past smokers had an RR of 2.0 (95% CI 1.2–3.4) for developing AMD compared with women who never smoked.5 There was a dose–response relationship between AMD and pack-years of smoking, and risk remained elevated for many years after smoking cessation. Results were consistent for various definitions of AMD, including wet AMD and dry AMD, with different levels of visual loss, and for different definitions of smoking. It was estimated that 29% of the AMD cases in that study could be attributable to smoking.5 These results were supported by a study among men participating in the Physicians’ Health Study.96 Several other studies have also shown an increased risk for AMD among smokers.65,97–99 Smoking is an important, independent, modifiable risk factor for AMD.
Mechanisms by which smoking may increase the risk of developing AMD include its adverse effect on blood lipids by decreasing levels of high-density lipoprotein (HDL) and increasing platelet aggregability and fibrinogen, increasing oxidative stress and lipid peroxidation, and reducing plasma levels of antioxidants.5 In animal models, nicotine has been shown to increase the size and severity of experimental CNV, suggesting that non-neuronal nicotinic receptors may also play a part in the effect of smoking on advanced AMD.100 Statistical interactions between smoking and either the Complement Factor H (CFH) Y402H or ARMS2/HTRA1 genotypes have not been confirmed (see “Genetics factors” below).101–102
The role of antioxidant vitamins in the pathogenesis of AMD has received a great deal of attention. Antioxidants, which include vitamin C (ascorbic acid), vitamin E (alpha-tocopherol), and the carotenoids (including alpha-carotene, beta-carotene, cryptoxanthin, lutein, and zeaxanthin), may be relevant to AMD because of their physiologic functions and the location of some of these nutrients in the retina. Lutein and zeaxanthin, in particular, are associated with macular pigment.103–105 Trace minerals such as zinc, selenium, copper, and manganese may also be involved in antioxidant functions of the retina. Antioxidants could prevent oxidative damage to the retina, which could in turn prevent development of AMD.10,106 Damage to retinal photoreceptor cells could be caused by photo-oxidation or by free radical-induced lipid peroxidation.107–108 This could lead to impaired function of the RPE and eventually to degeneration involving the macula. The deposit of oxidized compounds in healthy tissue may result in cell death because they are indigestible by cellular enzymes.108–109 Antioxidants may scavenge, decompose, or reduce the formation of harmful compounds.
The AREDS confirmed that antioxidant and zinc supplementation can decrease the risk of AMD progression and vision loss.4 This study included a double-blind clinical trial in 11 centers around the USA, randomly assigning 3640 participants to take daily oral supplements of antioxidants, zinc, antioxidants and zinc, or placebo. Both zinc alone and antioxidants and zinc together significantly reduced the odds of developing advanced AMD in participants with intermediate signs of AMD (see Chapter 65, Age-related macular degeneration: Non-neovascular early AMD, intermediate AMD, and geographic atrophy) in at least one eye. The zinc supplement included zinc (80 mg) as zinc oxide, and copper (2 mg) as cupric oxide; the antioxidant supplement included vitamin C (500 mg), vitamin E (400 IU), and beta-carotene (15 mg). If the AREDS formulation were used to treat the 8 million individuals in the USA who are at increased risk for developing advanced AMD, the AREDS group authors estimate that more than 300,000 would avoid advanced AMD and the associated vision loss during the next 5 years.2 AREDS-type supplements are a cost-effective way of reducing visual loss due to the progression of AMD.110 The effect of dietary antioxidants on the incidence of early AMD has not been established and there are questions about the effect of beta-carotene supplements on AMD since there is none in the retina, and high doses of zinc could have side-effects.
Diets high in antioxidant-rich fruits and vegetables may be related to a lower risk of exudative AMD. The first study launched to evaluate diet and AMD, the Dietary Intake Study, ancillary to the EDCCS, showed an inverse association between exudative AMD and dietary intake of carotenoids from foods.10 In that study reported in 1994, a diet rich in green leafy vegetables containing the carotenoids lutein and zeaxanthin was associated with a reduction in the risk of exudative AMD. Intake of 6 mg of lutein per day was associated with a significant 43% reduction in risk of AMD.10 A prospective double-masked study involving lutein and antioxidant supplementation in a group of 90 individuals showed that visual function was improved with 10 mg of lutein or a lutein/antioxidant formula.111 In a British study of 380 men and women, lower plasma levels of zeaxanthin were also found to be associated with an increased risk of AMD.112 A cross-sectional study using previously collected NHANES I data found a weak protective effect with increased consumption of fruits and vegetables rich in vitamin A.113 A prospective follow-up study has shown that fruit intake is inversely associated with exudative AMD. Participants who consumed three or more servings of fresh fruit per day have an RR of 0.64 (95% CI 0.44–0.93) compared to those who consumed less than 1.5 servings per day.7 The early evidence regarding lutein was recently supported by analyses of diet data from AREDS.114 AREDS2, an ongoing trial of lutein, zeaxanthin, and omega 3 fatty acids and assessment of omission of beta-carotene and use of much lower doses of zinc for the prevention of AMD progression, may provide additional data regarding optimal vitamin supplement regimens for AMD patients.
Studies that have examined the relationship between AMD and alcohol consumption have yielded mixed results. In the EDCCS, no significant relationship between alcohol intake and exudative AMD was noted in univariate analyses,71 but an inverse association could not be ruled out in multivariate analyses. Another case–control study found a suggestion of a nonlinear trend with higher risk of AMD in persons who had five drinks or more per day and a lower risk in persons who had one or two drinks per day compared with nondrinkers.115 In a case–control study using NHANES I data, moderate wine consumption was associated with a decreased risk of developing AMD, although the analysis did not control for the potential confounding effects of smoking.116 In a large prospective study, no support was found for a protective association between moderate alcohol consumption and risk of AMD, although there was a suggestion of a modest increased risk of AMD in heavier drinkers.117 The BDES found heavy drinkers were more likely to develop late AMD,98 whereas the BMES found an increased risk of early AMD only in current spirits drinkers.118 The evidence to date suggests that alcohol intake does not have a large effect on the development of AMD.
There is an association between AMD and overall obesity13,119–122 and abdominal adiposity.13 In a prospective cohort study of 261 individuals with some sign of nonadvanced AMD in at least one eye,13 individuals with a body mass index (BMI) between 25 and 29 had an RR of 2.32 (95% CI 1.32–4.07) for progression to advanced AMD when compared to those with a BMI of less than 25. Those with a BMI of at least 30 had an RR of 2.35 (95% CI 1.27–4.34) compared to the lowest category (BMI <25), after controlling for other factors. Similarly, the highest tertile of waist circumference had a twofold increased risk compared to the lowest tertile, and the highest tertile of waist-to-hip ratio had an RR of 1.84 compared to the lowest tertile. Thus, both overall and abdominal obesity were related to AMD progression. Vigorous physical activity three times a week reduced the risk of AMD progression by 25% compared to no physical activity.13 Obesity and physical activity are modifiable factors that may alter an individual’s risk of AMD incidence and progression. In one study, the susceptibility to advanced AMD associated with CFH Y402H was modified by body mass index (BMI), and both higher BMI and current and past smoking increased risk of advanced AMD within the same genotype category (see “Genetics factors” below).101
The literature to date regarding the association between sunlight exposure and AMD is conflicting. Overall, the data do not support a strong association between ultraviolet radiation exposure and risk of AMD, although a small effect cannot be ruled out. In the BDES,123 increased time spent outdoors in the summer was associated with a twofold increased risk of advanced AMD. The 5-year124 and 10-year125 incidence of early AMD in the BDES confirmed this association, although the 10-year incidence study showed few significant associations between environmental light and incidence and progression of early AMD. The EDCCS71 and the Pathologies Oculaires Liées à l’Age (POLA) study in France126 showed no significant association between advanced AMD and sunlight exposure. Sensitivity to sunburn may also be a risk factor. An Australian case–control study noted an association between sun-sensitive skin and risk of neovascular AMD.127 Conflicting results in these studies exemplify the difficulties encountered when studying this complex exposure. These include challenges in measuring acute and chronic lifetime exposure and the effect of potential confounding variables, such as sun sensitivity and sun avoidance behaviors. Furthermore, studies have evaluated different populations with different stages of AMD and people with varying intensity of exposures.