Racial Differences and Other Risk Factors for Incidence and Progression of Age-Related Macular Degeneration: Salisbury Eye Evaluation (SEE) Project

Abstract

purpose. To evaluate risk factors for the incidence and progression of age-related macular degeneration (AMD) in a racially heterogeneous, geriatric population.

methods. Subjects (n = 2240) aged 65 to 84 years underwent 2 examinations separated by 2 years, of which 1937 subjects (85%) were included in this report. Fundus photographs were performed at each examination and were graded by trained readers. Multivariate logistic regression models adjusted for age, sex, race, and clustering between eyes were used to evaluate risk factors for AMD incidence and progression.

results. Smoking was a strong, dose-dependent, risk factor for progression from medium size drusen to large drusen or pigmentary abnormalities within the central 1500-μm macular zone. Smoking was also a strong risk factor for development of incident focal pigmentation within 3000 μm of the foveal center. White participants were significantly more likely than blacks to develop large drusen and focal pigmentation and to progress from medium- to large-sized drusen or pigment abnormalities within the central 1500 μm macular zone. However, whites did not have an increased risk of progression from large drusen or pigment abnormalities within the central 1500-μm perimacular zone to foveal GA or CNV when compared with blacks.

conclusions. Smoking and race are important risk factors for progression from medium to large drusen or to pigment abnormalities within the central 1500-μm macular zone. Limitations in the power of this study preclude assessment of the roles of smoking and race on the ultimate progression to foveal GA or CNV once central large drusen or pigment abnormalities are present.

In developed nations such as the United States, age-related macular degeneration (AMD) is a leading cause of severe visual impairment. ¹ The prevalence of intermediate or unilateral advanced AMD is estimated at 8 million affected individuals in the United States alone. ² As the elderly fraction of the U.S. population grows, the absolute number of people affected with AMD is expected to increase. It has been predicted that nearly 3 million individuals will have geographic atrophy (GA) or choroidal neovascularization (CNV) by 2015. ²  

The exact causes of AMD are as yet unclear, but they are likely to be a combination of both genetic and environmental factors. Recent work has shown a significant relationship between AMD and a variation in the complement factor H (HF1) gene, ^{3 4 5} suggesting that inflammatory pathways may mediate the pathogenesis of AMD. Twin studies and ethnic variations in the incidence and prevalence of AMD implicate genetic factors in the pathogenesis of AMD; however, longitudinal studies of large cohorts suggest that modifiable factors such as smoking ^{6 7 8 9 10 11} and antioxidant/mineral usage, ¹² also affect AMD development or progression. Still other risk factors, such as sunlight exposure, ¹³ hypertension, ¹⁴ obesity, ^{10 15} hyperopia, ^{15 16 17} and cataract surgery, ¹⁸ have been shown inconsistently or inconclusively to be associated with AMD and may have to be studied further for their potential role in the pathogenesis of AMD to become clear. 

It can be difficult to determine whether risk factors identified within an ethnically homogeneous population are present due to genetic factors or from environmental exposure. In such studies as the Beaver Dam, ¹⁹ Blue Mountains, ⁸ Rotterdam, ⁷ and Barbados ²⁰ eye studies, relatively homogeneous populations composed of mostly white (Beaver Dam, Blue Mountains, and Rotterdam) or black (Barbados) participants have been examined. In contrast, the Salisbury Eye Evaluation (SEE) Project has the advantage of concentrating on an older cohort from a racially heterogeneous community that includes a significant percentage of black individuals, which may better allow distinctions to be made between racial groups. 

The purpose of this report is to present the incidence of AMD and specific fundus manifestations of AMD in the SEE Project and to explore risk factors, particularly racial differences, for incidence and progression of AMD in this cohort. Identification of risk factors may help elucidate the mechanisms and pathogenesis of AMD and potentially suggest strategies for intervention. In addition, individuals or groups at particular risk for disease progression may be more easily identified and eventually targeted for enhanced patient education, vision screening resources, and preventative interventions in the future. 

Methods

Beginning in July 1993, the SEE Project enrolled 2510 individuals aged 65 to 84 years who were identified through the Medicare database and resided in Salisbury, Maryland. Detailed methods of recruitment and study protocols have been published. ^{21 22 23} Briefly, in addition to detailed questionnaires regarding medical, ocular, and surgical histories; medication usage; and social habits, study subjects underwent a full ocular examination including visual acuity testing, lensometry/refraction, pupillary dilation, standardized retroillumination slit-lamp lens. and stereoscopic film-based fundus photography of fields 1 and 2. Participants were invited to return for a second full ocular examination, photography, and questionnaires approximately 2 years from the date of first participation. A total of 2240 subjects returned for the follow-up examination and were eligible for inclusion in this report. 

Fundus photograph grading was performed by two trained, independent readers who were masked to all demographic and ocular examination data. ²³ Discrepancies between the two readers were openly adjudicated, and unresolved differences were decided by a retina specialist (SBB). During the initial enrollment period, a set of 48 fundus photographs were recycled through the grading process to detect possible drift in the photograph grading. The weighted κ statistics were 0.72 or higher for each fundus characteristic graded when the first and last gradings were compared on this set of photographs. This process was not repeated for photographs from the follow-up examination, but the same two graders evaluated both the enrollment and the follow-up photographs. 

Photographs from the baseline and 2-year follow-up examinations were read independent of one another. Eyes with evidence of CNV or GA involving the foveal center were defined as having late AMD. Eyes with early AMD were classified as either early AMD 1 or early AMD 2. Early AMD 1 required the presence of at least one medium-sized druse (greatest linear dimension 64–125 μm) within 3000 μm of the foveal center in the absence of early AMD 2 or late AMD. Early AMD 2 required the presence of at least 1 large druse (>125 μm) or retinal pigment epithelium (RPE) abnormalities (hyperpigmentation or nongeographic atrophy of the retinal pigment epithelium) within the central macular zone (1500 μm of the foveal center), in the absence of late AMD. 

This protocol was approved in its entirety by the Institutional Review Board for the Johns Hopkins University School of Medicine. Written informed consent was obtained from all participants. This research adhered to the tenets of the Declaration of Helsinki. 

Inclusion and Exclusion Criteria

All subjects who had gradable photographs for AMD from both eyes at the baseline examination were included in this report. However, individuals with bilateral evidence of late AMD at the baseline examination were excluded, as they were no longer at risk of AMD progression. The fellow eyes of individuals with unilateral late AMD at baseline were also excluded, due to the known greater risk of progression to late AMD in these fellow eyes compared with the risk in eyes in individuals who do not have unilateral late AMD. ^{12 24 25} The rationale for the latter exclusion criterion is to omit the group of eyes that may indicate irreversible progression, potentially masking any other risk factors for AMD progression that may be important. Subjects were excluded if photographs of one or both eyes were not obtained at the baseline examination or were of inadequate quality to be graded, as the absence of late AMD in both eyes could not therefore be verified. 

Statistical Methods

The ages of subjects included and excluded from this analysis were compared by t-tests. Other demographic data were compared by using logistic regression, adjusting for age. 

All incidence rates are reported by eye, rather than by person. Incidence of AMD is reported in eyes with no AMD at baseline that had evidence of any AMD stage at follow-up (early AMD 1 or 2 or late AMD). Progression of AMD is reported for eyes with AMD at baseline that had a more severe stage of AMD at the follow-up examination (i.e., early AMD 1 progressing to early AMD 2). Incidence and progression of specific AMD manifestations such as central pigment abnormalities or large drusen were also evaluated. Outcome variables indicating specific types of progression of AMD were created. For example, when a variable was created that indicated progression from early AMD 1 to early AMD 2, only eyes with early AMD 1 at baseline were considered to be at risk for progression and were included in the referent group. Those eyes that were subsequently found to have early AMD 2 at follow-up were considered to have progressed. Progression variables were not mutually exclusive, i.e., eyes with early AMD 1 that progressed to early AMD 2 due to the appearance of central pigmentary abnormalities on follow-up examination were also considered to have progressed from no central pigment abnormalities to central pigment abnormalities. 

Incidence of AMD progression among white and black participants were compared by using the Fisher exact test and χ² tests. 

Risk factors for incidence or progression of AMD were identified by analyzing univariate relationships between dependent progression variables and independent variables such as race, sex, body mass index (BMI), visible light exposure, lens status, refractive error, history of smoking, hypertension, cardiovascular disease (congestive heart failure, angina, coronary artery disease), nutritional supplement usage, Mini Mental Status Examination, and education level. The definitions of each of these independent variables are described elsewhere. ^{21 22 23} Logistic regression models with generalized estimating equations (GEEs) were used to adjust for age and clustering between eyes. There were no adjustments for multiple testing. 

Independent variables related to AMD at P ≤ 0.10, adjusted for age and clustering between eyes, were then included in multivariate logistic regression models. To construct a parsimonious multivariate model, individual independent variables were then eliminated if no significant relationship was seen between the independent and dependent variable in the multivariate model and if elimination did not produce a significant change in the magnitude of the association between other independent variables and the dependent variable. 

Commercial software (Stata ver. 8.0; Stata Corp., College Station, TX) was used for all analyses. 

Results

Altogether, 2240 subjects completed the SEE examination at baseline and follow-up 2 years later, among which 1937 (86%) subjects are included in this report. The mean age of participants was 72.5 years; 59% were women, and 25% were black. A total of 303 subjects were excluded: 61 subjects due to late AMD at baseline in at least 1 eye and 242 subjects due to missing or ungradable photographs at baseline in at least 1 eye. Excluded subjects were more likely than included subjects to be older, black, and current or former smokers and to have a history of cataract surgery (Table 1 , column 2). Within the subgroup excluded due to known presence of late AMD at baseline in at least one eye, the subjects were more likely to be older and white, with a higher proportion of former or current smokers than included subjects. The subgroup of subjects excluded because of missing baseline photographs in at least 1 eye were more likely to be older, black, and free of arthritis and to have a history of cataract surgery and were less likely to have graduated from high school. 

Age-Adjusted Incidence and Progression of AMD by Race

Eyes with no AMD at baseline had a 2-year incidence of drusen at least medium in size within 3000 μm of the foveal center (early AMD 1) of 10.0%, of large drusen or pigment abnormalities within the central 1500 μm macular zone (early AMD 2) of 0.9%, and of late AMD of 0.3% (Table 2) . 

Progression from early AMD 1 to early AMD 2 during the 2-year interval occurred in 7.4%, progression from early AMD 1 to late AMD in 0.9%, and progression from AMD 2 to late AMD in 8.4%. Although the observed rates of progression were greater in whites than in blacks for each of these analyses, the differences were not statistically significant. 

A total of 3.5% of eyes that had at least medium-sized drusen at baseline progressed to large drusen, although this was more common in whites than blacks (4.1% vs. 1.6%, P = 0.01; Table 3 ). Progression from medium to large drusen specifically located in the central macula and large drusen to late AMD also occurred more frequently in whites than in blacks, although neither of these was statistically significant. 

Development of a substantial drusen area, as indicated by progression from drusen occupying less than 10% of the area encompassed by the entire grid template (diameter 6000 μm) to ≥10% was infrequent, as it hovered around 1% in each racial group. 

The 2-year incidence of focal hyperpigmentation within 3000 μm of the foveal center was 3.6% in whites and 1.6% in blacks (P = 0.006). Eyes with focal pigmentation at presentation within 3000 μm of the foveal center had progression rates of 13.6% in whites and 4.6% in blacks to late AMD. Eyes with large drusen (within 3000 μm of the foveal center) and either central RPE abnormalities (focal pigment or RPE depigmentation) or nonfoveal geographic atrophy within 3000 μm of the foveal center at baseline, the features identified by the Age-Related Eye Disease Study (AREDS) simple scale, as predictive of progression to late AMD, ²⁵ had a 2-year progression rate of 26.7% (12/45 eyes) to late AMD. 

In summary, race was a significant univariate risk factor (P < 0.05) for incident early AMD 1, incident focal pigmentation within 3000 μm of the foveal center, and progression from medium-sized drusen to large drusen. However, both races were equally likely to progress from no drusen to medium- or large-sized drusen. 

Multivariate Analysis of Risk Factors for Incident AMD or Progression of AMD

Whites were less prone to the development of incident early AMD 1 than were blacks (OR: 0.7 [0.50–0.99]; Table 4 ). However, whites were more likely to progress from early AMD 1 to early AMD 2 (OR: 2.1 [1.00–4.39]). In addition, older age, current smoking of at least one pack of cigarettes per day (compared with nonsmokers), and failure to complete high school education also had independent effects on increasing the risk of early AMD 1’s progressing to early AMD 2. Multivariate analyses that used smoking status (current/former/never) rather than number of cigarettes smoked per day as the smoking variable revealed that current smokers had 2.7 (95% CI 1.18–6.19) times the odds of progressing from early AMD 1 to early AMD 2, whereas former smokers did not have an increased risk of progression (data not shown in the table). 

Older age was the only nonocular factor found to be associated with a higher likelihood of progression from early AMD 2 to late AMD. A history of arthritis was associated with a lower odds (OR: 0.34 [0.13–0.89]) of progression from early AMD 2 to late AMD. Although whites appeared more likely to progress for many features of AMD development or progression, the small samples resulted in a lack of statistical significance. 

The odds of development of medium-sized drusen in eyes without medium-sized drusen at baseline was lower in whites than in blacks (OR: 0.69 [0.48–1.00]); however, progression from medium to large drusen was twice as common among whites as among blacks (OR: 2.25 [1.01–5.03]). 

The odds of development of RPE depigmentation or focal hyperpigmentation increased with advancing age. In addition, white race and smoking at least 10 cigarettes per day (compared with nonsmokers) were each associated with incident focal pigmentation. Multivariate analyses that used present smoking status (current/former/never) instead of number of cigarettes smoked per day as the smoking risk factor revealed that current smokers had higher odds of development of focal hyperpigmentation (OR: 1.9; 95% CI: 1.05–3.48) but that former smokers did not have an increased risk of progression (data not shown). 

Discussion

Multiple studies have explored potential relationships between various demographic, systemic, and ocular risk factors and incident cases of AMD and specific AMD fundus manifestations. ^{6 10 15 16 24 25 26 27 28 29 30 31 32} Although the Beaver Dam Eye Study has described progression of various AMD manifestations within their population, ^{31 33} risk factors for progression of specific AMD features have not been reported. 

Postulated risk factors for incident neovascular AMD include diabetes, ¹⁰ history of cataract surgery, ^{18 31} hypertension, ³³ hyperopia, ^{16 17} obesity, ¹⁰ diabetes, ^{28 29} and higher total serum cholesterol, ²⁹ but the role of these risk factors is unclear, as evidence in studies has been inconsistent. Risk factors for incident non-neovascular AMD, such as history of cataract surgery, ^{18 31} antacid use, ¹⁰ hyperopia, ^{16 17} visible light exposure, ³⁴ and higher total serum cholesterol ²⁹ have also been investigated, but results have varied among studies. 

However, the strongest and most consistent association of a modifiable risk factor and AMD has been smoking. Associations have been identified in multiple studies between smoking and development of both non-neovascular and neovascular AMD. ^{6 7 8 9 10 32} There are several theories that may explain the relationship between smoking and AMD. 

A theory regarding the pathogenesis of late AMD is that hypoxia stimulates the production of vascular endothelial growth factor (VEGF), which may lead to retinal endothelial cell proliferation and the development of neovascularization and neovascular AMD. ³⁵ Damage to blood vessels through promotion of atherosclerosis or vasoconstriction secondary to smoking ³⁶ may therefore further potentiate hypoxic retinal conditions, leading to an increased susceptibility to late AMD. It is less clear how hypoxic conditions may contribute to the pathogenesis and progression of early AMD. Some have found higher plasma VEGF levels in subjects with AMD compared with healthy control subjects, with comparable levels in subjects with non-neovascular and neovascular forms of the condition. ³⁷ This finding suggests that other factors in addition to VEGF contribute to the pathogenesis of AMD. 

Some have suggested that a major factor in the pathogenesis of AMD is oxidative stress leading to damage to the outer retina and RPE. ³⁸ Since it is known to decrease levels of circulating antioxidants, ³⁹ smoking may therefore decrease retinal defenses against oxidative damage. Smoking is also known to lead to activation of retinal phospholipase A2, which leads to production of inflammatory mediators such as leukotrienes and prostaglandins. ⁴⁰ Given that inflammation may play a major role in the pathogenesis of AMD, especially in those with a variation in the complement factor H gene (HF1), ^{3 4 5} a higher risk of AMD may therefore be found in smokers. 

We found a consistent relationship between smoking and progression of AMD in the SEE population. Current smokers appeared to have a dose-dependent increase in odds of progression from early AMD 1 (medium-sized drusen within the 3000-μm pericentral macular zone) to early AMD 2 (large drusen or pigment abnormalities within the 1500-μm central macular zone), with an odds ratio of 3.1 for those smoking at least a pack of cigarettes a day when compared with that of nonsmokers. Our data also support a relationship between smoking and incident AMD-related focal hyperpigmentation. Those who smoked at least a half a pack of cigarettes a day were twice as likely to develop focal pigmentation than were nonsmokers. We did not find a significant relationship between smoking and progression of AMD from a more advanced non-neovascular form (early AMD 2) to late AMD (foveal geographic atrophy or choroidal neovascularization), nor did we establish a relationship between smoking and the development of incident medium-sized drusen or progression from medium to large drusen. However, during this 2-year follow-up study, relatively few eyes progressed from early AMD 2 to late AMD (23/275) or from medium to large drusen (70/1991). It is possible that we failed to detect any relationship between smoking and these specific types of AMD progression because of the low event rates in our cohort. The strong relationship between smoking and incidence and progression of AMD supports the findings in previous studies, ^{6 7 8 9 10 32} and suggests that smoking cessation should be strongly encouraged in all patients with AMD. 

Participants who reported a history of arthritis were less apt to progress from early AMD 2 to late AMD. Specific characterization of type of arthritis (i.e., osteoarthritis or rheumatoid arthritis) was not performed in this cohort of subjects. The AREDS found that arthritis is associated with an increased risk of manifesting pigment abnormalities and intermediate or large drusen, ¹⁵ but this finding has not been demonstrated in other studies. The significance of the potentially protective effect of arthritis found in the present study may be difficult to interpret because of the relatively small number of eyes that progressed from early AMD 2 to late AMD. It is possible that arthritis is a marker for inflammation and that inflammation plays a role in development of the non-neovascular manifestations of AMD but a less significant role in the conversion of non-neovascular to neovascular disease. It is also possible that commonly used anti-inflammatory therapies for arthritis treatment such as corticosteroids and nonsteroidal anti-inflammatory medication may have decreased the amount of total inflammation in the body and thus decreased the risk of progression to late AMD. Completion of high school was associated with lower rates of progression from early AMD 1 to early AMD 2, which is consistent with findings from the AREDS, in which persons who failed to complete high school had higher rates of large drusen, CNV, or foveal geographic atrophy. ¹⁵ Similarly, completion of high school was found to be protective against development of neovascular AMD in the Eye Disease Case–Control Study. ⁴¹ It is unclear how level of education is biologically linked to AMD progression, and this maybe a surrogate risk factor related to unidentified confounding variables or confounding variables that have not been adequately controlled for. For instance, educational achievement is inversely associated with smoking. 

It has been observed that blacks have a lower prevalence of macular degeneration or lower prevalence of specific AMD features than do whites. ^{23 42 43} Some have hypothesized that the increased melanin in RPE cells of blacks may help act as a free radical scavenger or as a filter for ultraviolet radiation and may help protect the RPE cells and Bruch’s membrane, reducing the risk of development of large drusen and pigmentary changes. ¹⁵ However, direct comparisons of prevalence, incidence, and progression of AMD between racial groups in population-based cohorts have frequently been limited by the homogeneity of the populations studied. More recently, the Multi-Ethnic Study of Atherosclerosis (MESA) and the SEE project have evaluated cohorts that sampled different racial groups. The MESA found that the prevalence of AMD (lumping all stages of AMD) was significantly lower in blacks than in other racial groups ⁴³ and that the rates of late AMD did not differ significantly between blacks and whites. Prevalence rates of specific AMD features at entry into the SEE study were also consistent with lower rates in blacks than in whites for large drusen, drusen >250 μm, larger macular area involved with drusen, focal hyperpigmentation, and geographic atrophy. ²³  

In this report, we provide a direct comparison of the progression of AMD in blacks and whites in the SEE population and show by both univariate and multivariate analyses that whites were significantly more likely than blacks to develop new focal pigmentation within 3000 μm of the foveal center and to progress from medium to large drusen. Of note, we found that blacks were more likely than whites to develop incident early AMD 1 at the follow-up examination. It is possible that whites who were susceptible to development of AMD did so at an earlier age, and therefore AMD 1 was more likely to have already developed in whites by the time of the baseline examination than in their black counterparts. In addition, we did not find a significant racial association for progression from early AMD 1 to later phases of AMD or from early AMD 2 to late AMD, suggesting that race may not be an important risk factor for these specific progression rates. More likely, as the result of each analysis tended in the direction of lower risk for blacks, is that the limited sample size within a 2-year interval limited our power to detect such a difference. 

In the SEE population of black and white participants with an average age at entry of 73, we found that the 2-year rates of step-wise progression to each AMD category was approximately 10%. In 10% of those with no AMD at baseline, early AMD 1 developed in 2 years, whereas in 7.4%, it progressed from early AMD 1 to early AMD 2, and in 8.4%, from early AMD 2 to late AMD. Those with focal pigmentation had higher rates of progression to late AMD than those with large drusen (12.2% versus 6.0%); however, those eyes with both large drusen and central focal pigmentation had much higher rates of progression to late AMD (22.0%) than those with either one of these characteristics alone. Eyes with the highest risk for progression to late AMD at 2 years were those with large drusen (within 3000 μm of the foveal center) and either central RPE abnormalities (focal pigment or RPE depigmentation) or nonfoveal geographic atrophy within 3000 μm of the foveal center (26.7%). If the simplified severity scale as reported in AREDS Report No. 18 had been used, ²⁵ each of these eyes would have had a maximum eye severity score of 2. 

The strengths of this study include the relatively large proportion of African Americans evaluated in this population, as well as the relatively complete follow-up at the 2-year examination. Fellow eyes of those with documented unilateral late AMD at baseline were excluded to concentrate on nonocular risk factors for AMD progression. 

Limitations of this investigation stem from the relatively brief 2-year interval between examinations, which has led to a limited number of eyes progressing between AMD levels or developing incident AMD features. Although the proportion of blacks in this study is relatively high (>25%) in comparison to other population-based cohorts, the very limited number of eyes progressing to advanced AMD was insufficient to perform a separate risk factor analysis for blacks alone. In addition, multiple tests of significance were conducted to ascertain the relationship between risk factors and AMD. Repeated testing may therefore increase the likelihood that significant results, such as the associations between black race and incidence of early AMD signs and the protective effects of education status and history of arthritis, are the result of chance alone. However, finding associations between risk factors, such as smoking and race, that have been more consistent in previous studies with multiple types of AMD incidence and progression make these findings less likely to be attributable to chance alone. 

In conclusion, our data suggest that whites and heavy smokers are at higher risk of progression from medium to large-sized drusen or pigmentary abnormalities within the central 1500-μm macular zone. Limitations in the power of this study preclude assessment of the roles of smoking and race in the ultimate progression to foveal GA or CNV, once central large drusen or pigmentary abnormalities are present. 

Further study, including longer follow-up of this cohort, is needed to clarify racial differences in risk factors for AMD progression. 

 Supported by the Director’s Discovery Fund Award and Wilmer Research Grant (MAC) from the Wilmer Eye Institute; the Olga Keith Weiss Scholars Award (SBB) and a Senior Scientist Award (SKW) from Research to Prevent Blindness, an unrestricted grant from Alcon Research Institute (SKW), and Grant AG16294 from the National Institute on Aging (SKW).

 Submitted for publication December 11, 2007; revised January 25, 2008; accepted March 25, 2008.

 Disclosure: M.A. Chang, None; S.B. Bressler, None; B. Munoz, None; S.K. West, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Table 1.

 

Comparison of Included and Excluded Subjects

Variable	Included Subjects n (%)	Excluded Subjects n (%)	Excluded Due to Late AMD n (%)	Excluded Due to Missing or Ungradable Photographs n (%)
Subjects (n)	1937	303	61	242
Eyes (n)	3874	606	122	484
Age (y)
Mean ± SD	72.5 ± 4.8	74.5 ± 5.6*	76.1 ± 5.5*	74.2 ± 5.6*
Median	72.0	74.0	77.0	73.0
Range	65–86	66–86	66–86	66–86
Sex
Male	801 (41.4)	126 (41.6)	29 (47.5)	97 (40.1)
Female	1136 (58.7)	177 (58.4)	32 (52.5)	145 (59.9)
Race
White	1450 (74.9)	210 (69.3), †	53 (86.9), ‡	157 (64.9)*
Black	487 (25.1)	93 (30.7)	8 (13.1)	85 (35.1)
Smoking history
Never	774 (40.0)	109 (36.0), §	18 (29.5)	91 (37.6)
Past smoker	889 (46.0)	145 (47.9)	35 (57.4)	110 (45.5)
Current smoker	271 (14.0)	49 (16.2)	8 (13.1)	41 (16.9)
<10 cig. per day	60 (22.1)	15 (30.6)	1 (12.5)	14 (34.1)
10–19 cig. per day	82 (30.3)	15 (30.6)	1 (12.5)	14 (34.1)
≥20 cig. per day	122 (45.0)	17 (34.7)	5 (62.5)	12 (29.3)
Hypertension
Yes	1002 (52.0)	59 (53.5)	36 (60.0)	123 (51.9)
No	926 (48.0)	138 (46.7)	24 (40.0)	114 (48.1)
Diabetes
Yes	326 (16.9)	56 (18.5)	7 (11.5)	49 (20.3)
No	1610 (83.1)	246 (81.5)	54 (88.5)	192 (79.7)
Body mass index
<25	472 (24.7)	69 (24.1)	17 (28.3)	52 (23.0)
25–30	958 (50.2)	145 (50.7)	31 (51.7)	114 (50.4)
>30	479 (25.1)	72 (25.2)	12 (20.0)	60 (26.6)
Arthritis
Yes	1034 (53.4)	148 (48.8)	34 (55.7)	114 (47.1), ‡
No	903 (46.6)	155 (51.2)	27 (44.3)	128 (52.9)
Phakic status at baseline
Phakic eyes	3359 (87.1)	450 (78.5), †	96 (79.3)	354 (78.3), ¶
Pseudophakic eyes	497 (12.9)	123 (21.5)	25 (20.7)	98 (21.7)
Education
Did not graduate HS	958 (49.5)	170 (56.1)	29 (47.5)	141 (58.3), #
HS graduate	976 (50.5)	133 (43.9)	32 (52.5)	51 (41.7)

 Comparison tests conducted between included and excluded participants by using t-tests for age and logistic regression to adjust for age for all other variables. cig, cigarettes; HS, high school.

*   P < 0.001.

†   P = 0.004.

‡   P = 0.05.

§   P = 0.01.

∥   P = 0.03.

¶   P = 0.002.

#   P = 0.02.

Table 2.

 

Two-Year Incidence and Progression of AMD Levels by Race

Incidence or Progression Type	Yes	Total*	P , † (95% CI)
No AMD to early AMD 1
White	122 (9.1%)	1336	0.05
Black	55 (12.5%)	440	(0.50–0.99)
Total	177 (10.0%)	1776
No AMD to early AMD 2
White	10 (1.0%)	1029	0.69
Black	2 (0.7%)	282	(0.30–6.18)
Total	12 (0.9%)	1311
No AMD to late AMD
White	3 (0.3%)	1156	0.98
Black	1 (0.3%)	377	(0.10–9.27)
Total	4 (0.3%)	1533
Early AMD1 to early AMD 2
White	55 (8.3%)	664	0.13
Black	10 (4.7%)	212	(0.85–3.43)
Total	65 (7.4%)	876
Early AMD 1 to late AMD
White	8 (1.1%)	741	0.39
Black	1 (0.4%)	253	(0.32–18.9)
Total	9 (0.9%)	994
Early AMD 2 to late AMD
White	21 (8.7%)	242	0.72
Black	2 (6.1%)	33	(0.28–6.33)
Total	23 (8.4%)	275

 AMD definitions: early AMD 1, at least 1 medium-sized druse (greatest linear dimension 64–125 μm) within 3000 μm of the foveal center in the absence of early AMD 2 or late AMD; early AMD 2, at least 1 large druse (>125 μm) or hyperpigmentation or non-GA of the RPE within 1500 μm of the foveal center, in the absence of late AMD; late AMD, CNV or GA involving the foveal center.

*  Total number of eyes at risk differs between types of incidence and progression because not all photographs were gradable for all types of AMD manifestations.

†   P and 95% confidence interval (CI) from logistic regression for differences in proportions between whites and blacks, with adjustment for age and clustering between eyes.

Table 3.

 

Two-Year Incidence and Progression of Specific AMD Features by Race

Incidence or Progression Type	Yes	Total*	P , † (95% CI)
No drusen or small to medium drusen, ‡
White	108 (5.0%)	2176	0.05
Black	50 (7.0%)	712	(0.48–1.0)
Total	158 (5.5%)	2888
No drusen or small to large drusen
White	12 (0.7%)	1740	0.11
Black	3 (0.6%)	487	(0.22–5.27)
Total	15 (0.7%)	2227
Medium to large drusen, ‡
White	63 (4.1%)	1548	0.02
Black	7 (1.6%)	443	(1.14–5.63)
Total	70 (3.5%)	1991
Medium to large central drusen
White	31 (2.0%)	1545	0.09
Black	3 (0.7%)	442	(0.86–9.37)
Total	34 (1.7%)	1987
Large drusen to late AMD
White	16 (6.8%)	235	0.39
Black	1 (2.1%)	48	(0.30–21.4)
Total	17 (6.0%)	283
Drusen area ≤10% to >10%
White	5 (1.0%)	493	0.92
Black	1 (0.9%)	117	(0.13–9.25)
Total	6 (1.0%)	610
No focal pigmentation to focal pigmentation
White	93 (3.6%)	2581	0.006
Black	14 (1.6%)	861	(1.26–3.98)
Total	107 (3.1%)	3442
No RPE depigmentation to RPE depigmentation
White	43 (1.6%)	2651	0.97
Black	14 (1.6%)	873	(0.52–1.88)
Total	57 (1.6%)	3524
Focal pigmentation to late AMD
White	15 (13.6%)	110	0.33
Black	1 (4.8%)	21	(0.33–27.36)
Total	16 (12.2%)	131
Large drusen and central focal pigmentation to late AMD
White	8 (21.1%)	38	, §
Black	1 (33.3%)	3
Total	9 (22.0%)	41
Large drusen with central RPE abnormalities or nonfoveal GA to Late AMD
White	10 (24.4%)	41	, §
Black	2 (50.0%)	4
Total	12 (26.7%)	45

 Definitions of AMD are as in Table 2 . RPE abnormalities are focal pigmentation or RPE depigmentation.

*  Total number of eyes at risk differs between types of incidence and progression, because not all photographs were gradable for all types of AMD manifestations.

†   P and 95% confidence interval (CI) from logistic regression for difference in proportions between whites and blacks, with adjustment for age and clustering between eyes.

‡  Unless specifically mentioned, drusen and pigmentation characteristics refer to the area within the grid encompassing the area 3000 μm of the foveal center. Central refers to the area within 1500 μm of the foveal center. Drusen sizes: small, ≤63 μm; medium, 64–124 μm; large, ≥125 μm.

§  No comparisons were made between the two racial groups, since the number of eyes at risk was small.

Table 4.

 

Risk Factors for Different Types of AMD Progression

Progression Type	Odds Ratio	95% CI	P
No AMD to early AMD 1
Increasing age (per 5 y)	1.12	0.93–1.34	0.22
Male	0.95	0.69–1.30	0.73
White	0.70	0.50–0.99	0.05
Early AMD 1 to early AMD 2
Increasing age (per 5 years)	1.45	1.10–1.91	0.008
Male	1.48	0.87–2.50	0.15
White	2.10	1.00–4.39	0.05
Currently smoking <10 cigarettes per day*	0.87	0.11–6.74	0.89
Currently smoking 10–19 cigarettes per day*	2.13	0.63–7.17	0.22
Currently smoking ≥20 cigarettes per day*	3.07	1.19–7.94	0.02
Completion of high school education	0.57	0.33–0.97	0.04
Early AMD 2 to late AMD
Increasing age (per 5 y)	1.65	1.00–2.73	0.05
Male	0.49	0.16–1.55	0.23
White	1.18	0.23–6.14	0.85
History of arthritis	0.34	0.13–0.89	0.03
No drusen or small to medium drusen
Increasing age (per 5 y)	0.93	0.77–1.12	0.44
Male	1.02	0.74–1.43	0.88
White	0.69	0.48–1.00	0.05
Medium to large drusen
Increasing age (per 5 y)	1.6	1.21–2.10	0.001
Male	1.36	0.80–2.3	0.25
White	2.25	1.01–5.03	0.05
No RPE depigmentation to RPE depigmentation
Increasing age (per 5 y)	1.09	1.04–1.16	0.002
Male	1.66	0.96–2.87	0.07
White	0.96	0.51–1.83	0.91
No focal RPE hyperpigmentation to focal RPE hyperpigmentation
Increasing age (per 5 years)	1.28	1.04–1.59	0.02
Male	1.14	0.74–1.77	0.53
White	2.22	1.24–3.98	0.007
Currently smoking <10 cigarettes per day*	0.44	0.06–3.27	0.42
Currently smoking 10–19 cigarettes per day*	2.29	1.00–5.25	0.05
Currently smoking ≥20 cigarettes per day*	2.16	1.07–4.35	0.03

 Data are results of multivariate analyses adjusted for age, sex, race, and clustering between eyes. AMD definitions are as in Table 2 . Small drusen, ≤63 μm; medium drusen, 64–124 μm; large drusen: ≥125 μm.

*  Reference group is nonsmokers.