Prevalence of Age-Related Macular Degeneration in a Population-Based Sample of Hispanic People in Arizona: Proyecto VER

Objective  To report the prevalence of age-related macular degeneration (AMD) in a population-based sample of Hispanic individuals aged 50 years and older.

Methods  Proyecto VER (Vision and Eye Research) is a population-based study of blindness and visual impairment of Hispanic people in Arizona. Participants underwent complete ophthalmic evaluation, including stereoscopic fundus photography of fields 1, 2, and 4. All photographs for participants aged 50 years and older were graded using the Wisconsin Age-Related Maculopathy Grading system. The following signs were graded: drusen size, drusen type, and the area covered by drusen; pigmentary abnormalities; geographic atrophy; and exudative AMD.

Results  Sixty-seven percent (3178) of the original 4774 participants were 50 years of age or older. Of those, 92% (2928) had fundus photographs in at least 1 eye, and 95% (2780) of the photographs were of sufficient quality to grade early and late AMD.

Outcome Measures  The overall prevalence of late AMD was 0.5%. The prevalence increased from 0.1% in the 50- to 59-year age group to 4.3% in the group aged 80 years and older. Likewise, early AMD was strongly associated with age with a prevalence of 20% in the 50- to 59-year age group, increasing to 54% in the group aged 80 years and older. The prevalence of early AMD in Hispanic people was significantly higher than the reported prevalence in the white population. However, the prevalence of late AMD was lower than the estimates for the white population of the United States.

Conclusions  Although early macular changes were very common among Hispanic people, the prevalence of late AMD was infrequent. Further work is necessary to understand the underlying reasons for the different patterns of presentation of early and late signs of AMD among racial/ethnic groups and to characterize early AMD based on predictive value for severe disease in different populations.

Age-related macular degeneration (AMD) is the leading cause of blindness in older Americans,¹ affecting white people more than black people.^2-6 Although Hispanic people are the second largest minority in the United States, there are limited data available on the prevalence of AMD in this ethnic group.

The third National Health and Nutrition Examination Survey^5,6 reported significantly lower prevalence of late AMD in Hispanic individuals than in non-Hispanic white individuals, based on relatively small actual numbers of Mexican American individuals. In that nationwide survey, macular changes were graded based on nonstereoscopic fundus photographs taken through a nondilated pupil of only 1 eye. This assessment methodology is likely to underestimate the prevalence of all AMD signs, and comparisons of racial/ethnic groups could be biased. Data from the third National Health and Nutrition Examination Survey showed substantial differences by ethnic group, which varied by age. In the Colorado-Wisconsin Study of Age-Related Maculopathy,^7,8 stereoscopic fundus photographs in 731 Hispanic people were graded similarly to photographs from 3999 non-Hispanic white people from the Beaver Dam Eye Study. Incidents of late-stage AMD were significantly less frequent among Hispanic people than non-Hispanic white people, but there was no ethnic difference in the prevalence of any AMD. The lower prevalence of late AMD in Hispanic people was not explained by differences in the presence of any of the risk factors examined. The relatively small numbers in the Hispanic group did make prevalence estimates unstable.

Precise prevalence estimates by racial/ethnic group, as well as additional research to gain understanding of the underlying reasons for the different patterns of presentation of early and late signs of AMD among racial/ethnic groups, is important in developing strategies for prevention and treatment of AMD.

In Proyecto VER (Vision and Eye Research), we had the unique opportunity to determine the prevalence of AMD in a large, population-based sample of Mexican American people living in Arizona. The purposes of this article are first to report the prevalence of early and late signs of AMD in a population-based sample of Hispanic individuals and, second, to compare the prevalence with age-specific estimates for white and black people in the United States.

Proyecto VER is a population-based study of blindness and visual impairment in Mexican American individuals living in Nogales and Tucson, Arizona. Details of sample selection and participation rates are provided elsewhere.⁹ In summary, a stratified random sample was used to select block groups in 2 major cities in Arizona: Nogales in Santa Cruz county and Tucson in Pima county. All members of sampled households who were self-described Hispanic and aged 40 years and older were eligible to participate. Overall, 72% of eligible subjects were part of the study. Data collection started in April 1997 and ended in September 1999. In this study, photographs were graded only for 3178 participants aged 50 years and older.

After we obtained written informed consent, participants were interviewed at home and invited for a clinic visit at a central site. All study procedures were approved by the Johns Hopkins University (Baltimore, Md) Joint Committee on Clinical Investigations and followed the tenets of the Declaration of Helsinki.

The home questionnaire contained specific questions on education, smoking history, and self-reported diabetes status and a series of questions to assess Native American ancestry. Blood pressure was taken using a standard protocol; specifically, we took 3 separate readings of systolic and diastolic pressure and used the average of the second and third measures. Hypertension was defined as having a systolic blood pressure of 160 mm Hg or greater, having a diastolic blood pressure of 90 mm Hg or greater, or receiving treatment for hypertension. Blood was drawn to determine the level of glycosylated hemoglobin; these procedures have been described previously.⁹ In this paper, diabetes was defined as either giving an affirmative answer to the question of whether diabetes had been diagnosed by a physician or having an glycosylated-hemoglobin level of 7.0% or greater. All others were classified as not having diabetes.

A standardized dilated ophthalmic examination was performed on all participants by an ophthalmologist (J.R.) to assign causes of vision loss and assess ocular health.

Stereoscopic fundus photographs of Diabetic Retinopathy Study standard fields 1, 2, and 4 were taken through dilated pupils using a Zeiss 30° fundus camera (Carl Zeiss, Oberkochen, Germany). The photographs were read by trained graders at the University of Wisconsin Ocular Epidemiology Reading Center, Madison. A modification of the Wisconsin Age-Related Macular Grading System^10,11 was used. The grading involved a preliminary and detailed grading followed by a photograph edit and adjudication if necessary. The presence of the signs described next was graded for each eye.

For drusen, 3 features were considered: size, area, and type. The maximum drusen size was graded with 1 of the following categories: none, questionable, less than 63 μm in diameter, between 63 and 125 μm in diameter, between 126 and 249 μm in diameter, 250 μm or greater in diameter, or reticular drusen. The drusen area was estimated using all drusen present in field 2 and assigned to 1 of the following categories: none, questionable, occupying an area of a circle with a diameter of less than 63 μm, between 63 and 104 μm in diameter, between 105 and 249 μm in diameter, between 250 and 499 μm in diameter, between 500 and 749 μm in diameter, between 750 and 1499 μm in diameter, and 1500 μm or greater in diameter. The area covered by drusen was estimated using the area of a circle with a diameter of 63 μm for the smallest area category and an area midway between the lower and upper boundary for subsequent categories. The ratio of the estimated affected area to the total area of the grid was calculated to obtain the percentage of area covered by drusen. We graded the maximum drusen type, using the most severe drusen present in field 2, as none, questionable, hard distinct, soft distinct, soft indistinct, or reticular drusen.

In gauging retinal pigment epithelial (RPE) abnormalities, we considered 2 aspects, increased retinal pigment and RPE depigmentation. Increased retinal pigment was graded as absent, questionable, total area of hyperpigment occupying an area of a circle with a diameter of less than 63 μm, between 63 and 125 μm in diameter, greater than 125 μm but less than 6.3% of the subfield, or 6.3% of the subfield or greater. Retinal pigment epithelial depigmentation was graded as absent, questionable, total area of RPE depigmentation occupying a circle with a diameter of less than 125 μm, between 125 μm and less than 6.3% of the subfield, between 6.3% of the subfield and one half of the disc area, between one half of the disc area and less than 1 disc area, or 1 disc area or greater.

We considered 5 types of late AMD lesions, geographic atrophy, pigment epithelial detachment, detachment of sensory retina, subretinal hemorrhage, and subretinal disciform scar.

The following definitions were used in the analysis. Exudative AMD was considered present if any of the following lesions were present: pigment epithelial detachment, detachment of sensory retina, subretinal hemorrhage, or subretinal disciform scar. Late AMD was present if there were signs of exudative AMD or pure geographic atrophy. Early AMD was present if late AMD was absent and there were signs of soft indistinct or reticular drusen or any drusen (except hard indistinct) combined with RPE changes in the macular area.

Eyes were considered gradable if they could be graded for drusen size. If at least 1 eye had evidence of a specific sign, the person was considered to have that sign.

Intergrader and intragrader agreement was assessed using the quadratic weighted κ statistic on a random subset of 30 eyes. There was excellent interobserver and intraobserver agreement on the presence of early AMD, exudative AMD, and geographic atrophy (κ, 0.8-1.0). There was substantial interobserver and intraobserver agreement on the presence of retinal pigmentary changes (κ, 0.6-0.9) and moderate agreement on drusen size and type (κ, 0.4-0.8)

We present the prevalence of early and late AMD, neovascular AMD, geographic atrophy, drusen characteristics, and RPE abnormalities stratified by age and sex. Logistic regression models were used to examine the relationship between the main outcomes and sex, adjusting for age. Odds ratios and 95% confidence intervals are presented. We used age-stratified estimates of late AMD and drusen greater than 125 μm for white and black people in the US population as presented by Friedman et al¹² and the Proyecto VER population structure as the standard population to calculate age-adjusted prevalences. We used our estimates of late AMD and drusen greater than 125 μm for comparison purposes.

Overall, 4774 or 72% of eligible Hispanic individuals participated in the study. Characteristics of nonparticipants have been described previously.⁹ Briefly, nonparticipants were more likely to be men, to belong to the younger and older age groups, and to report fewer problems with their vision. There were no differences in self-report of diabetes. In this analysis, we include the 3178 participants (66.3% of the original sample) who were 50 years of age or older at the time of the clinic examination. The mean age of participants with gradable photographs was 68 years, ranging from 50 to 96 years. Almost one quarter were 70 years of age or older, half were current or past smokers, close to half had hypertension, and 25% had diabetes.

Fundus photographs of at least 1 eye were obtained for 2928 subjects (92% of the selected sample), and 95% of the photographs were of sufficient quality to be able to grade late signs of AMD and drusen size (Figure). The main reasons for not having photographs taken were inability to have the pupil dilated, significant lens opacity, and participant refusal. Participants not included in this analysis because the photographs were not available or were of insufficient quality to grade drusen characteristics were significantly older, more likely to have had cataract surgery, more likely to be visually impaired, more likely to report diabetes, and less likely to have 12 or more years of education (Table 1).

The distribution of drusen by size was age dependent with the size increasing with age. Similar distribution was observed for men and women across all age groups (Table 2). Drusen less than 63 μm in diameter as the largest sized drusen was the most common sign, with prevalence decreasing from 63% in the younger ages to 31% in the older age group as large drusen became more predominant. Presence of large drusen (>125 μm in diameter) were fairly common with the prevalence increasing from 11% in the younger age group to almost 40% in the group aged 80 years and older (test for trend, P<.001). In addition, the proportion of the large drusen that was greater than 250 μm increased with age from 15% in the younger age group to 37% after age 80 years.

The prevalence of soft drusen, distinct or indistinct, significantly increased with age, from 19% in the 50- to 59-year age group to 50% in the group aged 80 years and older (test for trend, P<.001). Similar prevalence of soft drusen was observed for men and women (age-adjusted, P = .83), and one third of soft drusen was indistinct (Table 3).

Overall, 92% of the population had at least 1 drusen in at least 1 eye (Table 4). The mean percentage area covered by drusen in the macula increased with age from 0.18% in the youngest age group to 1.93% in the group aged 80 years or older (P<.001). After adjusting for age, there was no significant difference in the percentage area covered by drusen between men and women (age-adjusted, P = .07).

Similar prevalences of RPE depigmentation and increased retinal pigment were observed. Overall, 3% of the population was affected with each sign (Table 5). The prevalence of pigmentary abnormalities was less than 4% in the younger age groups, and after age 70 years, there was a rapid increase from 6% in the 70- to 79-year age group to over 12% in people aged 80 years or older (test for trend, P<.001). Pigmentary abnormalities were more likely to occur in men than in women, with an age-adjusted odds ratio (95% confidence interval) of 1.87 (1.29-2.70). There was a strong correlation between the presence of pigmentary abnormalities and the size and type of drusen (data not shown); more than 70% of the pigmentary abnormalities occurred in eyes with drusen greater than 125 μm in diameter and with soft indistinct drusen.

Occurrence of late AMD was infrequent in this population. Both signs of exudative AMD and geographic atrophy increased with age (Table 6). Geographic atrophy was more prevalent than exudative AMD. Overall, only 15 people had signs of late AMD. Pure geographic atrophy was bilateral in 4 persons and unilateral in 7 persons; exudative AMD was bilateral in 2 persons; and 2 persons had signs of both geographic atrophy in one eye and exudative AMD in the other, both of whom were in the oldest age group. Early AMD was present in approximately one third of the population (Table 6). The prevalence of early AMD was age dependent, increasing from 20% in the younger age group to over 50% after age 80 years (test for trend, P<.001). Similar prevalence was observed for men and woman across all ages (age-adjusted, P = .76).

The age-sex specific prevalence rate of late AMD and drusen greater than 125 μm in diameter for non-Hispanic white people and black people has been estimated from pooled population-based studies.¹² We compared the observed prevalence in our Hispanic population with the estimated age-specific prevalence of AMD in white and black people in the United States. Although the prevalence of drusen greater than 125 μm in diameter for the Proyecto VER population was consistently higher than the prevalence in the white population, late AMD was substantially lower (Table 7). In the younger age groups, the prevalence of late AMD was lower than for white and black people, but in the group aged 80 years and older, the observed prevalence in Hispanic individuals was higher than in black individuals and substantially lower than in white individuals. The age-adjusted prevalence of late AMD, using Proyecto VER as standard population, were 0.58%, 1.75%, and 0.96% for Hispanic, white, and black people, respectively. In contrast, the prevalence of drusen greater than 125 μm in diameter was higher than for black or white people across all ages. The corresponding age-adjusted prevalences were 17.8%, 7.6%, and 6.5% for Hispanic, white, and black people, respectively.

Some persons without photographs were judged by the ophthalmologist to have late AMD. Using the correlation between diagnosis in the clinic and photograph assessment of exudative disease and geographic atrophy, we expected to have 2 cases of late AMD in people who did not have photographs. The inclusion of these 2 additional cases did not change substantially the prevalence of late AMD in Hispanic individuals.

Data from Proyecto VER provide prevalence estimates of early and late macular changes in a large population-based sample of Mexican American individuals. The prevalence estimates for early AMD or drusen greater than 125 μm in diameter appear to be much higher than what have been reported for non-Hispanic white individuals and black individuals in the United States¹² and in the Beaver Dam Eye Study, a population-based study using identical grading.¹³ However, the prevalence of late AMD was substantially lower than the prevalence estimates for the non-Hispanic white US population, including estimates from the Beaver Dam Eye Study. Such differences have been suggested by previous findings in the third National Health and Nutrition Examination Survey^5,6 and in the Colorado-Wisconsin Study of Age-Related Maculopathy.⁸

The observed differences in frequency of AMD in Hispanic people and non-Hispanic white people are unlikely the result of differentially excluding from analysis the people more likely to be affected with late AMD in our study. Overall, 11% (371) of Proyecto VER participants were excluded from late AMD prevalence calculations because photographs were not available. For the prevalence of late AMD in Hispanic people to reach the level of what has been estimated for white people, 40 out of the 371 nonparticipants (or 11%) needed to have had late AMD, which is a very unlikely scenario. To further explore the magnitude of the possible underestimation, we examined the clinical assessment of AMD in both groups (with and without photographs). We estimated that we were likely to be missing 2 additional cases of late AMD. After including those cases, the prevalence of late AMD was still substantially lower than the prevalence reported for non-Hispanic white people.

For Hispanic people in Proyecto VER, the ratio of geographic atrophy to exudative macular degeneration was 3.2. This is similar to ratios found in Icelandic white people¹⁴(4.8) and lower than those found in Inuit people in Greenland¹⁵(12.0). It is higher than the ratios found in most other studies, especially in studies of white people in Beaver Dam, Wisconsin¹²(0.5); in the Blue Mountains¹⁶(0.4) and Victoria¹⁷(0.7), Australia; and in Rotterdam,¹⁸ the Netherlands (0.9), all of which used a similar grading system to detect AMD. The reason for this variation in the prevalence of both late stages of AMD is not known.⁷

It could be argued that the lower rates of late AMD in Hispanic people are due to lower survival rates in those with late signs. However, the association with AMD and mortality has not been consistent. In Beaver Dam, there was no relation, but in 2 other studies,^19,20 severe AMD has been associated with higher mortality. In 1 study, the excess mortality has been partially explained by history of cardiovascular disease. However, if mortality is explaining the lower rates of severe AMD in Hispanic people, the differential mortality would have to be higher for Hispanic than for non-Hispanic white people. In this cross-sectional study, we cannot exclude that possibility.

Another possibility is that Hispanic people may not progress to geographic atrophy or exudative AMD, despite the higher prevalence of early AMD signs.^21-28 Reasons for this are unknown. Choroidal and RPE pigmentation have been postulated to be protective against progression to neovascular AMD. Genetic differences have also been postulated to explain these differences.^29-32 It does appear to be inconsistent to have a high prevalence of early AMD and a low prevalence of late AMD, although the Baltimore Eye Survey also reported similar findings in African American people.² The question of whether large drusen are more likely to progress to advanced AMD in this population can only be addressed in longitudinal studies. Such studies in different racial and ethnic groups are warranted to further characterize the natural history of AMD.

Correspondence: Beatriz Muñoz, MSc, Wilmer Eye Institute, Room 118, Johns Hopkins University, 600 N Wolfe St, Baltimore, MD 21287 (bmunoz@jhmi.edu).

Submitted for Publication: November 10, 2003; final revision received November 18, 2004; accepted January 5, 2005.

Financial Disclosure: None.

Funding/Support: This study was supported by Public Health Service research grant EY13783 from the National Eye Institute, Bethesda, Md. We are grateful to the team of Proyecto VER for their data collection and to the University of Wisconsin Ocular Epidemiology Reading Center, Madison, for the grading of the photographs. Dr West is a senior scientific investigator for the Research to Prevent Blindness, New York, NY.