Prevalence of Age-Related Macular Degeneration in the United States

Objective  To estimate the prevalence and distribution of age-related macular degeneration(AMD) in the United States by age, race/ethnicity, and gender.

Methods  Summary prevalence estimates of drusen 125 µm or larger, neovascularAMD, and geographic atrophy were prepared separately for black and white personsin 5-year age intervals starting at 40 years. The estimated rates were basedon a meta-analysis of recent population-based studies in the United States,Australia, and Europe. These rates were applied to 2000 US Census data andto projected US population figures for 2020 to estimate the number of theUS population with drusen and AMD.

Results  The overall prevalence of neovascular AMD and/or geographic atrophyin the US population 40 years and older is estimated to be 1.47% (95% confidenceinterval, 1.38%-1.55%), with 1.75 million citizens having AMD. The prevalenceof AMD increased dramatically with age, with more than 15% of the white womenolder than 80 years having neovascular AMD and/or geographic atrophy. Morethan 7 million individuals had drusen measuring 125 µm or larger andwere, therefore, at substantial risk of developing AMD. Owing to the rapidlyaging population, the number of persons having AMD will increase by 50% to2.95 million in 2020. Age-related macular degeneration was far more prevalentamong white than among black persons.

Conclusion  Age-related macular degeneration affects more than 1.75 million individualsin the United States. Owing to the rapid aging of the US population, thisnumber will increase to almost 3 million by 2020.

Age-related macular degeneration (AMD) is the leading cause of blindnessamong European-descended people older than 65 years.^1-6 Whileindividual studies have arrived at estimates of both early AMD and more severestages of the disease (advanced AMD), the numbers of persons identified withAMD, and the numbers of individuals in the oldest age groups are frequentlysmall, yielding somewhat unstable estimates.

Recent research has demonstrated that the likelihood of vision lossfrom AMD can be reduced in individuals with high-dose vitamin supplementation⁷ and can be reduced with laser photocoagulation andphotodynamic therapy in others with neovascular (NV) forms of the disease.^8-11 However,the reduction in visual loss using these treatments for persons with geographicatrophy (GA) and for most persons with NV AMD is limited. Policy plannersrequire solid estimates of disease prevalence to determine the likely benefitof these and future therapies. This study was undertaken to determine moreaccurately the prevalence of intermediate and advanced AMD in the United Statesby pooling findings from large population-based studies that have been conductedover the last 2 decades and applying the prevalence rates to the US population.Furthermore, we have estimated the expected increase in AMD in the future.

We focused on obtaining accurate estimates of the prevalence of NV AMDand GA as well as the presence of large drusen (ie, those ≥125 µm).Given the large overlap of pigmentary changes with large drusen, and datathat show that large drusen are associated with an almost 6% risk of developingadvanced AMD over 5 years in the involved eye (Ronald Klein, MD, MPH, writtencommunication, September 16, 2002), and because it had been measured in allof the populations, we chose to look at large drusen as a manifestation ofintermediate AMD. Therefore, the definitions used for this study are as follows.

Age-related macular degeneration is defined as that named AMD by theInternational ARM Study Group, and includes the following 2 groups¹²: (1) those with GA, which is a discrete area of retinaldepigmentation at least 175 µm in diameter with a sharp border and visiblechoroidal vessels in the absence of NV AMD in the same eye; and (2) thosewith NV AMD, which is serous or hemorrhagic detachment of either the retinalpigment epithelium or sensory retina, the presence of subretinal fibrous tissue,or minimal subretinal fibrosis and widespread retinal pigment epithelial atrophy.Or it is any AMD that represents the presence of GA or NV AMD in either eye.

Subjects were classified as having GA if either eye had GA. Similarly,subjects were classified as having NV AMD if either eye had NV, so some individualswere counted in both categories of AMD. Therefore, estimates for any AMD willbe lower than the combined total of estimates for NV AMD and GA.

Large drusen were defined as drusen 125 µm or larger in diameterin the macula (defined as a region 3000 µm in diameter centered on thefoveola) in either or both eyes.

To be included in the present research, the studies contributing datahad to (1) be population based, (2) provide data on the conditions being studied;and (3) use a standard photographic grading system for determining the prevalenceof AMD and drusen. Since these studies were conducted when no preventive treatmentswere available for AMD, all known studies meeting these criteria (both publishedand unpublished) were included. These include the Baltimore Eye Survey, Baltimore,Md¹³; the Barbados Eye Study, Barbados, WestIndies¹⁴; the Beaver Dam Eye Study, BeaverDam, Wis¹⁵; the Blue Mountains Eye Study, Sydney,New South Wales, Australia¹⁶; the RotterdamStudy, Rotterdam, the Netherlands¹⁷; the MelbourneVision Impairment Project, Melbourne, Victoria, Australia¹⁸;and the Salisbury Eye Evaluation Project, Salisbury, Md¹⁹ (Table 1). The Baltimore Eye Survey enrolled5308 black and white persons (75% of the intended population), 19% of whomhad no gradable photographs for AMD.¹³ TheBarbados Eye Study enrolled 4314 black persons (84% of the intended population),17% of whom had no photographs gradable for AMD.¹⁴ TheBeaver Dam Eye Study enrolled 4926 subjects (83% of the intended population,4896 of whom were white), 2% of whom had no gradable photographs for AMD.¹⁵ The Blue Mountains Eye Study enrolled 3654 whitepersons (82% of the eligible population), 2% of whom had no gradable photographsfor AMD.¹⁶ The Rotterdam Study had 6872 personsvisit the study center (67% of the intended population), 0.9% of whom hadno gradable photographs for AMD.¹⁷ The SalisburyEye Evaluation Project enrolled 65% of the eligible population (all personswere ≥65 years), 2% of whom had no gradable photographs.¹⁹ TheMelbourne Vision Impairment Project enrolled 4345 (92% of the eligible population)of 4744 persons.¹⁸

The investigators from each study provided us with the number of individualswith gradable photographs in at least 1 eye and the number found to have NVAMD, GA, or any AMD in at least 1 eye, and at least 1 large druse in 1 orboth eyes stratified by gender and race for groups aged 40 through 44, 45through 49, 50 through 54, 55 through 59, 60 through 64, 65 through 69, 70through 74, 75 through 79, 80 through 84, 85 through 89 years, and 90 yearsand older.

The age-specific prevalence proportions were derived in 2 steps. First,pooled prevalence proportions were estimated for each race/ethnic-, gender-,and age-specific stratum using minimum variance linear estimation. Stratum-specificproportions from each study were transformed using a logarithm odds transformation,and proportion variances were estimated assuming the binomial distribution.Second, logistic regression models were fit to the pooled prevalence proportionsusing the midpoint of each age interval as the independent variable. Modelswere fit separately by gender for GA, NV AMD, any AMD, and the presence ofdrusen in white persons. Prevalence rates for drusen in the Beaver Dam EyeStudy were significantly higher (Figure 1) than those found in the other studies and, thus, were excludedfrom the pooled prevalence estimates for drusen in white persons.

Age and race/ethnicity effects in the models were tested using the Wald χ² test statistic. Odds ratios for gender differences were based on Mantel-Haenszel χ² tests for the 2 × 2 tables of observed rates, adjusting forage and the study effect.

Owing to the few cases of AMD in black persons, models were only fitfor any AMD and drusen. Age-specific prevalence rates for NV AMD and GA inblack persons were estimated by applying a constant factor to the modeledage-specific rates for any AMD. These factors (0.73 for NV AMD and 0.27 forGA in women, and 0.59 for NV AMD and 0.41 for GA in men) were based on theobserved proportion of all individuals with AMD who were identified with eitherGA or NV AMD in either eye. Prevalence rates for NV AMD, GA, any AMD, anddrusen in Hispanic persons and other races/ethnicities were assumed to besimilar to rates found in black persons.

The number of cases of AMD in the United States in each race/ethnicity,gender, and age category were generated by applying the modeled prevalencerate for each year of age to the 2000 US Census population and summing acrossthe age range for each 5-year age category. Projected estimates were derivedin the same manner, using US Census middle-series population projections for2020. Stratum-specific US prevalence rates were computed by dividing the totalnumber of estimated cases for each stratum by the stratum-specific US population.Estimates for AMD in Western Europe and Australia were based on applying theage- and gender-specific rates found in white persons to their respectivepopulations 40 years and older.

Pooled data for European-descended individuals from the Baltimore EyeSurvey, the Blue Mountains Eye Study, the Beaver Dam Eye Study, the RotterdamStudy, the Melbourne Vision Impairment Project, and for those older than 65years from the Salisbury Eye Evaluation Project found a strong age-relatedincrease in the prevalence of large drusen, GA, NV AMD, and any AMD (Figure 1 and Figure 2). There were dramatic increases in rates for both men andwomen older than 80 years, with models showing a highly significant quadraticterm for age. Of white females in the 50- through 54-year-old range, 0.20%had AMD vs 1.52% of those aged 70 through 74 years and 16.39% of those olderthan 80 years (Table 2). Geographicatrophy was slightly less prevalent than NV AMD in all age groups. While largedrusen were present in more than 1% of those aged 40 through 49 years, nocases of AMD were found in any of the studies for white persons in this agegroup. Rates were similar between men and women (age-adjusted odds ratio [OR]for men = 1.01; 95% confidence interval [CI], 0.81-1.25).

Pooled data for black persons were derived from 3 studies—theBarbados Eye Study, the Baltimore Eye Survey, and the Salisbury Eye EvaluationProject. Drusen prevalence was strongly age related, while the prevalenceof AMD increased less dramatically with age (Table 2). Black women aged 50 through 54 years had a 0.68% prevalence of AMD, which increased to 1.47% for those aged 70 through 74 yearsand to 2.44% for those aged 80 years and older. Five individuals younger than50 years (all cases of NV AMD from the Barbados Eye Study) were identifiedamong the 1705 evaluated, even though drusen were present in more than 2%of those in this age range. Although women had higher rates of AMD than men,differences were not statistically significant (age-adjusted OR for men =0.64; 95% CI, 0.31-1.32).

Applying age-, race/ethnicity-, and gender-specific rates to the USpopulation as determined in the 2000 US Census, we estimate that 1.75 millionindividuals (1.47%; 95% CI, 1.38-1.55%; Table 3) have AMD, 1.22 million (1.02%; 95% CI, 0.93-1.11%; Table 4) have NV AMD in at least 1 eye,973 000 individuals (0.81%; 95% CI, 0.77-0.86%; Table 5) have GA in at least 1 eye, and 7.3 million individuals(6.12%; 95% CI, 5.93-6.31%; Table 6)have large drusen (ie, ≥125 µm in diameter) in either 1 or both eyes(approximately 50% of these individuals have bilateral large drusen). Thesenumbers are estimated to increase substantially in the coming decades. Thenumber of individuals in the United States with AMD is estimated to increasemore than 50% from 1.75 million in 2000 to 2.95 million in 2020. Applyingthe age-, race/ethnicity-, and gender-specific rates, Australia is estimatedat present to have 130 000 cases of AMD; Western Europe has 3.35 millioncases of AMD. Applying the higher rates for the prevalence of large drusenfound in the Beaver Dam Eye Study, we would expect an additional 6.4 millionwhite individuals in the United States to have large drusen in either 1 orboth eyes.

From pooled data from population-based eye diseases prevalence studies,we estimate that, at present, 1.75 million individuals in the United Stateshave either GA or NV AMD in at least 1 eye. Approximately 7.3 million personsare estimated to have the early stages of AMD defined by large retinal drusenin at least 1 eye. Individuals with at least 1 druse of 125 µm or largerin diameter had a 5.8% risk of progressing to advanced AMD in the same eyeover 5 years in the Beaver Dam Eye Study (Ronald Klein, MD, PhD, written communication,September 16, 2002), indicating that they are, indeed, at high risk. Similarrates were recently published in the Rotterdam Study in which eyes with fewerthan 10 drusen 125 µm or larger in diameter had a 5-year risk of GAof 2% and of NV AMD of 1%. Those with 10 drusen or more had a 14% risk forboth GA and NV AMD.²⁰ We estimate that 2.95million persons in the United States will have AMD in 2020. It is unknownwhat the effect of high-dose vitamin supplementation using a combination ofzinc, beta-carotene, and vitamins C and E will be on these projected rates.⁷

Age-related macular degneration is associated with severe reductionsin quality of life as documented by clinical depression in almost one thirdof those who have this condition.²¹ Among individualswho have vision loss from AMD, 60% report significant declines in their abilityto participate in valued activities.²² Greateraccess to and use of rehabilitative services for those with irretrievablevision loss will be needed as advanced AMD affects a growing number of olderAmericans.

The current research has several limitations. First, although this isa meta-analysis of population-based studies, none of the studies enrolledall eligible subjects. On average, about 20% of those eligible did not participate,which may cause bias in the estimates. Nonparticipants may include more individualswith known disease, as these persons may not see any benefit to participating.Conversely, nonparticipants may have had better ocular health and did notparticipate because they saw no value in receiving a free eye examination.

A second limitation is the lack of gradable photographs. The numberof ungradable photographs was higher in the 3 studies of black persons. Thecameras used were different in those studies. While this higher rate of ungradablephotographs may be attributable to cataract, it is also possible that thosewith poor central vision were harder to image, which would have led to underestimatesin these populations. The Baltimore Eye Survey documented that those who wereblind from AMD were less likely to have gradable photographs.¹³

A third important limitation is the lack of data on other minority USpopulations. For black persons the data in the younger ages were limited,and the numbers studied in the older age ranges were not particularly large.To date there are few data that provide stable estimates of AMD prevalencein Hispanic persons. In the National Health and Examination Survey (NHANES)III, rates were based on 1 photograph of 1 eye of each subject. Age-relatedmacular degeneration was present in 0.5% of non-Hispanic white persons, 0.13%of non-Hispanic black persons, and 0.06% of Mexican Americans 40 years orolder.²³ In the San Luis Valley Study, SanLuis Valley, Colo, of the 571 Hispanic persons aged 43 through 74 years livingin Colorado sampled for the study, the prevalence of early maculopathy (pigmentarychanges and/or soft drusen) was 10%, slightly lower than that found for whitepersons (14%, P<.05), but AMD was found in onlyperson.²⁴ Both of these studies suggest a lowrate of AMD in Hispanic persons compared with white and black persons. Basedon these articles, it is possible that our estimates of AMD in Hispanic personsand those of other races/ethnicities, which assume the same prevalence ratesas black persons, may overestimate the number of individuals with AMD in thesepopulations. However, a recent report from Arizona near the Mexican borderfound that AMD was the second leading cause of visual impairment in that HispanicAmerican population, indicating that we may have underestimated AMD prevalencein this population.²⁵ These estimates willneed to be updated as prevalence surveys are published in these populations.Caution must be taken in using our data to compare rates between racial/ethnicgroups.

Fourth, differential misclassification may affect the estimated ratesof AMD in black persons. For example, it is possible that eyes with polypoidalchoroidopathy may be incorrectly classified as NV AMD in black persons comparedwith white persons, a possible explanation for the higher rates in youngerblacks compared with younger whites.

This article gives the best available estimate for the magnitude ofthe problem of AMD in the United States based on a meta-analysis of population-baseddata. The number of US population affected by AMD is increasing as the populationages. More than 1 in 10 white individuals 80 years and older has advancedAMD. A determined effort to identify effective preventive strategies willbe needed if we are to avoid a large increase in the numbers of persons havingthis condition.

Corresponding author: David S. Friedman, MD, MPH, Wilmer Eye Institute,Wilmer 120, 600 N Wolfe St, Baltimore, MD 21287 (e-mail: david.friedman@jhu.edu).

Submitted for publication June 26, 2003; final revision received October31, 2003; accepted December 8, 2003.

From the Dana Center for Preventive Ophthalmology, Wilmer Eye Institute,Johns Hopkins University, Baltimore, Md (Drs Friedman, Muñoz, Kempen,and Congdon); Department of Epidemiology and Biostatistics, School of PublicHealth and Health Services, George Washington University Medical Center, Washington,DC (Ms O'Colmain); Macro International, Inc, Calverton, Md (Ms O'Colmain);Department of Ophthalmology and Visual Science, University of Wisconsin, Madison(Ms Tomany); the Centre for Eye Research Australia, University of Melbourne,East Melbourne, Victoria (Ms McCarty); Ophthalmogenetics, The NetherlandsOphthalmic Research Institute KNAW, Amsterdam (Dr de Jong); Department ofEpidemiology and Biostatistics, Erasmus Medical Center, Rotterdam, the Netherlands(Dr de Jong); Department of Ophthalmology, Academic Medical Center, Amsterdam,the Netherlands (Dr de Jong); Department of Preventive Medicine, School ofMedicine, Stony Brook University, Stony Brook, NY (Dr Nemesure); and Departmentof Ophthalmology, University of Sydney, Sydney, New South Wales, Australia;Centre for Vision Research, Westmead Hospital, Westmead, New South Wales,Australia (Dr Mitchell).

Dr Friedman is supported in part by the Robert E. McCormick Award fromResearch to Prevent Blindness Inc, New York, NY.

Article

The members of The Eye Diseases Prevalence Research Group are as follows: The Baltimore Eye Survey, Baltimore, Md: James M. Tielsch;Alfred Sommer; Joanne Katz; David S. Friedman; Harry A. Quigley. The Barbados Eye Studies, Barbados, West Indies: M. Cristina Leske;Suh-Yuh Wu; Barbara Nemesure; Anselim Hennis; Leslie Hyman; Andrew Schachat. Beaver Dam Eye Study, Beaver Dam, Wis: Barbara E. K. Klein;Ronald Klein; Kristine E. Lee; Scot E. Moss; Sandra C. Tomany. Blue Mountains Eye Study, Sydney, New South Wales, Australia: PaulMitchell; Jie Jin Wang; Elena Rochtchina; Wayne Smith; Robert G. Cumming;Karin Attebo; Jai Panchapakesan; Suriya Foran. The MelbourneVisual Impairment Project, Melbourne, Victoria, Australia: Hugh R.Taylor; Cathy McCarty; Bickol Mukesh; LeAnn M. Weih; Patricia M. Livingston;Mylan Van Newkirk; Cara L. Fu; Peter Dimitrov; Matthew Wensor. Proyecto VER (Vision Evaluation Research), Nogales and Tucson, Ariz: SheilaWest; Beatriz Muñoz; Jorge Rodriguez (deceased); Aimee Broman; DanielFinklestein; Robert Snyder. Rotterdam Study, Rotterdam,the Netherlands: Paulus T. V. M. de Jong; Johannes R. Vingerling; RogerC. W. Wolfs; Caroline C. W. Klaver; Albert Hofman; Redmer van Leeuwen; M.Kamran Ikram; Simone de Voogd. Salisbury Eye EvaluationProject, Salisbury, Md: Sheila West; Gary Rubin; Karen Bandeen Roche;Beatriz Muñoz; Kathy Turano; Oliver D. Schein; Donald Duncan.