Location of Lesions Associated with Age-Related Maculopathy Over a 10-year Period: The Beaver Dam Eye Study

Abstract

purpose. To describe cumulative incidence and changes in retinal lesions associated with age-related maculopathy (ARM) by location over a 10-year period and to examine the relation of location of those lesions to progression of ARM.

methods. Persons ranging in age from 43 to 84 years and living in Beaver Dam, Wisconsin, at the time of a census (1987–1988) were examined two to three times over a 10-year period (n = 3684). Drusen area, size, and type; retinal pigment epithelium depigmentation; increased pigment; geographic atrophy; and neovascular macular degeneration were determined in each of nine macular subfields: central, inner and outer superior, inner and outer nasal, inner and outer inferior, and inner and outer temporal by grading of stereoscopic color fundus photographs. Late ARM was defined as presence of either geographic atrophy or neovascular ARM.

results. Lesions were more likely to change or develop in specific locations. Drusen area increased most in the central circle. Compared with other quadrants, drusen greater than 125 μm in diameter and soft indistinct or reticular drusen were most likely to develop in the superior or temporal quadrants, whereas pigmentary abnormalities were most likely to occur in the nasal or superior quadrants. In general, large drusen, soft indistinct drusen, and pigmentary abnormalities were more likely to develop in the inner circle versus the central and outer circles. The quadrant location of early ARM lesions in 72 persons in whom late ARM developed was generally similar to that in persons who did not have late ARM. However, persons who had geographic atrophy were more likely to have large drusen in the inner circle than in the outer circle, while those who did not have late ARM were more likely to have large drusen in the outer circle.

conclusions. Lesions associated with early ARM were more likely to develop in specific locations in the macular area, and persons with lesions closer to the fovea may be related to a higher risk of development of late ARM. The quadrant location of early ARM lesions does not appear to add additional information to the risk of development of late ARM.

Late stage age-related maculopathy (ARM), defined by either presence of geographic atrophy or neovascular ARM, is a leading cause of blindness in the Western world among people aged 65 years and older. ¹ There are few population-based data regarding the distribution in the macular area of early lesions associated with ARM (e.g., drusen and pigmentary abnormalities) and their incidence and relationship to late stages of the disease. Understanding of the relation of location of the incidence and progression of early stages of ARM may result in a better understanding of the pathogenesis of the disease. For example, differences in the incidence of large soft drusen or pigmentary abnormalities might suggest a vascular, anatomic, or environmental exposure as being involved in the disease. Distribution of large amounts of drusen in the perifoveal area might also enable better risk prediction of progression to late stages of the disease, such as geographic atrophy. In a previous report, we evaluated the prevalence of these early lesions by location within the macular area at the baseline examination of the Beaver Dam Eye Study. ² We found that the lesions did not distribute uniformly throughout the macular area. To investigate further, in the present study, we evaluated the location of such lesions at 5- and 10-year follow-up examinations of the same cohort. We specifically attempted to answer the following questions: (1) Within different locations of the macula, how does drusen area change? (2) Where in the macula are new ARM lesions likely to develop? (3) For eyes in which late ARM develops, is there a particular location compared with others that have a specific lesion present before the late lesion occurs? 

Methods

Population

Methods used to identify and describe the Beaver Dam cohort have appeared in previous reports. ^{3 4 5 6 7} In brief, a private census of the population of Beaver Dam, Wisconsin, was performed from September 15, 1987, to May 4, 1988, to identify all residents in the city or township of Beaver Dam who were 43 to 84 years of age. Of the 5924 eligible individuals, 4926 (83%) participated in the baseline examination between March 1, 1988 and September 14, 1990. ³ Ninety-nine percent of the population was white. Of those surviving, 3684 (81.1%) participated in the 5-year follow-up examination between March 1, 1993, and June 14, 1995. Of the 3334 surviving participants in the baseline and second examination, 2764 (82.9%) participated in the second follow-up examination between March 1, 1998, and June 9, 2000. All data were collected with institutional review board approval in conformity with all federal and state laws, and the study was in compliance with the tenets of the Declaration of Helsinki. The mean and median times between the baseline and 10-year follow-up examinations were 10.1 years and 10.0 years, respectively, and the SD was 0.4 years. Comparisons between participants and nonparticipants at the time of the baseline and the 5- and 10-year follow-up examinations have appeared elsewhere. ^{4 5} In general, live nonparticipants were older and had more medical conditions. After adjustment for age and gender, persons with early ARM lesions were as likely to participate as persons without early ARM lesions (data not shown). 

Procedures and Definitions

At each examination, stereoscopic 30° color fundus photographs centered on the disc (Diabetic Retinopathy Study standard field 1) and macula (Diabetic Retinopathy Study standard field 2) and a nonstereoscopic color fundus photograph temporal to, but including, the fovea of each eye were taken. ^{6 8} Before grading, a grid consisting of three circles concentric with the center of the macula and two perpendicular diagonal crosshairs was superimposed over one member of the stereoscopic pair of field 2 (Fig. 1) . An independent member of the grading team applied the grids to the photographs so that the subfields corresponded to the same location for each visit. The radius of the innermost circle corresponded to 500 μm in the fundus of an average eye, and the radii of the middle and outer circles corresponded to 1500 and 3000 μm, respectively. Nine subfields were defined by the grid: central, inner and outer superior, inner and outer nasal, inner and outer inferior, and inner and outer temporal. The four inner subfields together compose the inner region; the four outer subfields compose the outer region. Inner and outer subfields were combined to form four quadrants: superior, nasal, inferior, and temporal. The ratio of the areas corresponding to the central subfield, the inner region, and the outer region was 1:8:27. The ratio of the area of the central subfield to that of any of the four quadrants was 1:8.75. 

Two gradings were performed for each eye. Graders were masked to any characteristics (e.g., age, sex, medical conditions) of the persons whose eyes they were grading. First, a preliminary grading was performed by one of two senior graders. Next, detailed gradings were performed by one of three other experienced graders. For detailed grading, each eye was assessed independently of the fellow eye. The assessment consisted of a subfield-by-subfield, lesion-by-lesion evaluation of each set using the Wisconsin Age-Related Maculopathy Grading System. ^{9 10} Next, a series of edits and reviews was performed. The detailed grades of the presence and severity of specific lesions were compared with that of the preliminary grading. Standardized edit rules were used to adjudicate disagreements. ¹⁰  

Finally, the detailed graders were asked to make side-by-side comparisons between baseline, 5-year, and 10-year follow-up photographs of eyes that showed change for ARM lesions between baseline and follow-up. ¹¹ These edits were done in random order of visits. 

Grading included evaluation of drusen area, presence of drusen larger than 125 μm in diameter, presence of soft indistinct or reticular drusen, presence of retinal pigment epithelial (RPE) depigmentation, and presence of increased pigment. More detailed descriptions of these lesions appear elsewhere. ¹² Drusen areas were estimated for each subfield in which hard distinct, soft distinct, or soft indistinct drusen were present. Drusen area for eyes without drusen was defined as 0 μm². Geographic atrophy was defined by the presence of one or more sharply defined, usually more or less circular patches of complete depigmentation of the RPE, which typically expose choroidal blood vessels. A patch must be 175 μm or more in diameter anywhere in the macular area to be considered. Neovascular ARM was defined as any part of the macular area having an RPE detachment and/or a subretinal hemorrhage and/or a subretinal fibrous scar and/or treatment for any of these late neovascular lesions. Eyes with either geographic atrophy and/or neovascular ARM were defined as having late ARM. 

Statistical Analysis

A variety of statistical methods were used. The change in drusen area from baseline to 10-year follow-up was calculated, and a repeated-measures test of equality was used to examine uniformity of mean drusen area change among the different locations. ¹³ We evaluated 10-year cumulative incidence rates by location, using the product limit method. ¹⁴ To test for differences in incidence (or prevalence in the late ARM subgroup analysis) between different locations, we used the generalized estimating equation (GEE) approach. ¹⁵ We excluded eyes that had the particular lesion in any location at baseline. Each eye contributed three observations to the circle analyses and four observations to the quadrant analyses. It should be noted that lesions could be incident in more than one location. The GEE methodology controls for the correlation between locations within the same eye. ¹⁵ The subsequent probability (P) testing for a difference between locations was produced from the type III test of this analysis. To compensate for the wide age range of the population, all models were adjusted for age. All analyses were performed on computer (SAS ver. 8.1; SAS, Cary, NC). ¹⁶  

For incidence of new lesions and area change, analyses were performed separately for right and left eyes. Results were similar for the eyes. For economy of space, results are presented for right eyes only. If late ARM was present at a visit, then all early lesion data in that eye at that visit was set to missing. In a subgroup analysis of persons in whom late ARM developed, we used data from that eye. For persons who had late ARM in both eyes, we chose the right eye. Similar analyses were performed with the left eye, and results were similar. There were three persons in whom geographic atrophy developed by the 5-year follow-up, and neovascular ARM developed in that same eye by the 10-year follow-up. These persons were considered to have geographic atrophy. 

Results

Question 1: Within Different Locations of the Macula, How Does Drusen Area Change?

In each subfield evaluated, drusen area increased significantly over time (P < 0.001). Further, the increase in area was larger in persons who were older at baseline (P < 0.001). Figure 2 shows age-adjusted changes in drusen area for each particular subfield of interest. Because some subfields have different overall areas than others, all subfield areas were rescaled to the same scale as the central circle. Drusen area did not change uniformly across the macular area (P < 0.001 for the three circles and for the four quadrants). For example, the superior and temporal quadrants had larger increases in drusen area than the nasal and inferior quadrants. After rescaling other subfields to the central circle, the central circle had the largest increase in drusen area over 10 years. 

Question 2: Where in the Macula Are New ARM Lesions Likely to Develop?

We evaluated the 10-year cumulative incidence of lesions defining early ARM, large drusen greater than 125 μm in diameter (Fig. 3) , soft indistinct or reticular drusen (Fig. 4) , RPE depigmentation (Fig. 5) , and increased retinal pigment (Fig. 6) by age groups for each location in the macular area. Incidence of all lesions increased with age in all locations (P < 0.001). Further, there were significant differences between incidence rates at different locations in the macular area. 

Compared with other quadrants, large drusen greater than 125 μm in diameter were least likely to develop in the inferior quadrant and compared with other circles, least likely to develop in the center circle. Compared with other quadrants, soft indistinct or reticular drusen were most likely to develop in the temporal or superior quadrant and, compared with other circles, were least likely to develop in the center circle. Compared with other quadrants, RPE depigmentation was most likely to develop in the superior or nasal quadrant and, compared with other circles, was least likely to develop in the outer circle. Compared with other quadrants, increased pigment was most likely to develop in the nasal quadrant. It was equally likely to develop in all circles (P = 0.24). These associations were not altered by including the late ARM status of the fellow eye (data not shown). 

Question 3: For Eyes in Which Late ARM Develops (Neovascular ARM or Geographic Atrophy), Is There a Particular Location Compared with Others That Have a Specific Lesion Present before the Late Lesion Occurs?

Late-stage ARM developed in at least one eye of 72 (2.0%) persons by the 10-year follow-up examination. Forty-one eyes had neovascular ARM (two persons had geographic atrophy in the fellow eye). Thirty-three eyes had geographic atrophy (two persons had neovascular ARM in the fellow eye). In general, early ARM lesions were likely to be in more than one quadrant at baseline in eyes in which late ARM lesions developed (Table 1) . Further, early lesions had slightly higher rates of being in more than one quadrant for geographic atrophy than for neovascular ARM. In general, in persons who had neovascular ARM and/or geographic atrophy, the inner circle was more likely to have early lesions than the central or outer circle. In addition, the prevalence rates of early lesions in the inner circle were higher in geographic atrophy than in neovascular ARM. In contrast, there was not a consistent location distribution within the quadrants, and these patterns were not much different between neovascular ARM and geographic atrophy. Eyes in which neovascular macular degeneration developed were more likely to have had large drusen in the temporal and superior quadrants than in the nasal and inferior quadrants. Eyes in which geographic atrophy developed were more likely to have large drusen in the temporal quadrant than in the other three quadrants. Results were similar for soft indistinct or reticular drusen (Table 1) . There was very little evidence of a difference in distribution in quadrants of pigmentary abnormalities in preceding development of late ARM (P > 0.20 for both neovascular ARM and geographic atrophy). 

Table 1 also presents the prevalence rates of early lesions for all persons in whom late ARM did not develop. The rates were much lower, but the location distribution was generally similar. For example, persons who did not have late ARM were more likely to have had soft indistinct drusen in the superior or temporal quadrant. Similarly, persons in whom either neovascular ARM or geographic atrophy developed were more likely to have had soft indistinct drusen in the superior or temporal quadrant. There were a few instances in which the location pattern was different. For example, persons who did not have late ARM were more likely to have had large drusen in the outer circle, whereas those who had geographic atrophy were more likely to have had large drusen in the inner circle. Because of the low incidence of late ARM (∼2%), we could not evaluate significance of these differences in location. 

Discussion

One of the original goals of the Beaver Dam Eye Study was to examine the association of location of early ARM lesions in the macular area to the incidence and progression of disease to further understanding of the possible pathophysiologic mechanisms associated with this disease. For this reason, the grading system used three concentric circular grids with two perpendicular diagonal crosshairs to divide the macular area into quadrantic and circular areas. We had hypothesized that early ARM lesions would be found in different locations, reflecting possible environmental (light) and genetic influences and that the presence of these lesions in specific locations would be associated with increased risk of progression to signs of late ARM. In a previous report, we found that the various lesions associated with early ARM are located in specific patterns in the macular area. ² This study extends these findings from that study to look at incidence and progression of these early lesions over a 10-year period. 

In a previous report, we evaluated the relationship between early lesions to the progression of disease. ¹¹ We found that incidence rates of geographic atrophy and neovascular ARM were much higher in persons with large drusen, soft indistinct drusen, RPE depigmentation, or increased retinal pigment at baseline. An important finding of our present study was evidence that persons in whom late ARM developed were more likely to have early lesions in the inner circle (500 μm from the fovea). Differences between the inner and outer circles were larger with geographic atrophy than with neovascular ARM. This finding is consistent with observations from Sarks et al. ^{17 18} who reported the concentric appearance of drusen and pigmentary abnormalities outside the foveal area and their convergence toward the fovea as geographic atrophy develops. We found that persons who did not have geographic atrophy were more likely to have RPE depigmentation or increased pigment in the inner circle. There are few comparative data from population-based epidemiologic studies. Data from the Blue Mountains Eye Study contains information regarding location from the fovea using similar concentric circles as used in the Beaver Dam Eye Study. The investigators reported that the closer drusen were to the foveal center at baseline, the higher the likelihood of development of incident geographic atrophy or neovascular macular degeneration 5 years later. ¹⁹ They did not measure location by quadrants in that study. 

The study provides further insights regarding progression of early ARM lesions by location. With respect to other quadrants, the largest increase in area of drusen and the location where large drusen or soft indistinct drusen developed was in the superior and temporal quadrants. This was consistent with our earlier finding of the highest frequency of large soft drusen within the superior and temporal quadrants. In contrast, pigmentary abnormalities were more likely to develop in the nasal quadrant. It is not known why lesions develop in these locations and whether they represent genetic, postnatal, environmental, anatomic, or physiologic influences. Our data suggest a possible familial association of RPE depigmentation. ²⁰ We found this association to be strongest in the nasal quadrant, but the data may have been underpowered to detect a possible familial association in other locations (Knudtson MD et al., unpublished data, 2003). There are no anatomic nasal–temporal asymmetries in the sensory retina that explain the differences in location. Differences in blood flow within the choriocapillaris may, in part, explain the differences. Blood flow within the choriocapillaris has been reported to be more interconnected and less lobular nasally than temporally. ²¹ This results in fewer watershed zones nasally. Thus, it is possible that the incident nasal–temporal drusen asymmetries we observed reflect a choroidal blood flow asymmetry and that the nasal–temporal RPE depigmentation asymmetry reflects a genetic influence. These possibilities remain speculative. Finally, others have speculated that factors such as response of the lids to light may affect different areas of the retina differentially, possibly accounting for location differences of incident early ARM lesions (e.g., higher incidence of soft indistinct drusen in superior macular compared with inferior macular quadrants). ²²  

Strengths of our study include a large community-based population, a high response rate at both baseline and follow-up examinations, and use of standardized assessment of risk factors and end points. However, any conclusions or explanations regarding associations described herein must be made with caution. First, the analyses conducted for this report included the examination of many associations. As is the case in all studies, some statistically significant associations may be due to chance when, in fact, no association exists. Second, the incidence of late ARM end points was less than 2%, limiting our ability to describe specific location differences in lesions between persons who do and do not have late-developing ARM. 

In summary, lesions associated with ARM were more likely to develop in specific locations in the macular area, and early ARM lesions located closer to the fovea may be related to the development of late ARM. The reasons for this remain unknown. The quadrant location of early ARM lesions does not appear to add additional information to the risk of development of late ARM. More precise location data of specific lesions, using computer-assisted approaches in large cohorts may provide further clarification of location and its association to the incidence and progression of ARM. 

 Supported by National Eye Institute Grant EY06594 (RK, BEKK).

 Submitted for publication October 1, 2003; revised January 23 and March 10, 2004; accepted March 12, 2004.

 Disclosure: M.D. Knudtson, None; R. Klein, None; B.E.K. Klein, None; K.E. Lee, None; S.M. Meuer, None; S.C. Tomany, 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.

 

Figure 1.

 

Grid defining nine subfields in the macular region of a right eye: CC, center circle; IS, inner superior; IN, inner nasal; II, inner inferior; IT, inner temporal; OS, outer superior; ON, outer nasal; OI, outer inferior; OT, outer temporal.

Figure 2.

 

Age-adjusted 10-year rescaled mean drusen area changes by location in right eyes. Areas have been divided appropriately so that the mean changes are displayed on a scale comparable to the central circle: the ratio between central to inner to outer circle is 1:8:27; the ratio of the central circle to any quadrant is 1:8.75.

Figure 3.

 

Right eye 10-year cumulative incidence of drusen more than 125 μm in diameter, by age. Data are stratified by (a) circles and (b) quadrants. There were significant differences in incidence rates (adjusted for age) between circles (P < 0.001) and between quadrants (P < 0.001), based on GEE models.

Figure 4.

 

Right eye 10-year cumulative incidence of soft indistinct or reticular drusen by age. Data are stratified by (a) circles and (b) quadrants. There were significant differences in incidence rates (adjusted for age) between circles (P < 0.001) and between quadrants (P < 0.001), based on GEE models.

Figure 5.

 

Right eye 10-year cumulative incidence of RPE depigmentation by age. Data are stratified by (a) circles and (b) quadrants. There were significant differences in incidence rates (adjusted for age) between circles (P < 0.001) and between quadrants (P < 0.001) based on GEE models.

Figure 6.

 

Right eye 10-year cumulative incidence of increased pigment by age. Data are stratified by (a) circles and (b) quadrants. There were no significant differences in incidence rates (adjusted for age) between circles (P = 0.24), but there were significant differences between quadrants (P < 0.001), based on GEE models.

Table 1.

 

Percentage with Specific Early ARM Lesions at Baseline by Macular Location Comparing Persons that Have Neovascular ARM or Geographic Atrophy in at least 1 Eye to those that Do Not Have Late ARM

ARM	n	Circles					P *	Quadrants							P *	Quad, † (% >1)
				CC	IC	OC			SUP	NAS	INF	TEM
Drusen >125 μm in diameter
Neovascular	41	36	58	58	<0.01	58	33	38	53	0.02	56
Geographic	33	48	82	69	<0.01	53	52	53	76	0.09	73
Either, ‡	72	41	69	63	<0.001	56	41	44	63	0.01	64
No late, §	3316	2	4	6	<0.001	4	3	3	4	<0.001	39
Soft indistinct or reticular drusen
Neovascular	41	36	58	50	0.01	54	33	43	50	0.04	49
Geographic	33	24	73	66	<0.001	66	48	41	73	0.03	73
Either, ‡	72	30	65	57	<0.001	59	39	41	61	<0.01	60
No late, §	3316	1	4	5	<0.001	3	2	2	4	<0.001	45
RPE depigmentation
Neovascular	41	15	30	15	0.04	23	23	15	20	0.44	24
Geographic	33	33	48	9	<0.01	36	30	27	24	0.47	33
Either, ‡	72	23	38	13	<0.001	28	27	20	21	0.38	28
No late, §	3316	1	2	0	<0.001	1	1	1	1	<0.001	35
Increased pigment
Neovascular	41	20	43	18	<0.01	23	33	15	25	0.28	27
Geographic	33	52	64	18	<0.001	39	48	33	36	0.27	39
Either, ‡	72	34	52	18	<0.001	30	39	23	30	0.05	32
No late, §	3316	4	4	3	<0.001	2	3	2	2	<0.001	26

 Data are expressed as percentages. Abbreviations are defined in Figure 1 .

*  Age-adjusted P testing for difference between subfields based on generalized estimating equation models.

†  Percentage of persons with a lesion in more than one quadrant; for the no-late-AMD category, this was calculated based on the number of persons with at least one quadrant prevalent at baseline.

‡  Two persons had neovascular ARM in one eye and geographic atrophy in the other. Data from the right eyes was used for these two persons (1 neovascular, 1 geographic).

§  Persons known to not have late ARM at follow-up in either eye. Data are shown for right eyes.

The authors thank Rick Chappell (Department of Biostatistics, University of Wisconsin-Madison) for consulting on the statistical analysis and Tien Wong (Department of Ophthalmology, University of Singapore) for comments and suggestions.