Download Iris pigmentation and extent of disease in patients with neovascular

Iris Pigmentation and Extent of Disease in Patients With
Neovascular Age-Related Macular Degeneration
Michael A. Sandberg, Alexander R. Gaudio, Sumiko Miller, and Asher Weiner
Purpose. To determine whether (he extent of disease in age-related macular degeneration
(AMD) varies with iris pigmentation.
Methods. The authors assessed visual function and macular appearance in the fellow eye or
both eyes of 132 white patients with unilateral neovascular AMD. All patients had a visual
acuity of 20/(50 or better in the fellow eye. Eighty-nine of the patients were coded as having
light irides (blue, green, or hazel) and 43 were coded as having dark irides (brown); the two
groups of patients had comparable mean ages.
Results. By the Mann-Whitney test for differences in mean rank, fellow eyes with light irides
showed a marginally worse visual acuity (P = .156) but significantly more visual field impairment by letter recognition perimetry (P = .011) and the macular threshold test of the
Humphrey Field Analyzer (P - .043), and more retinal pigment epithelial atrophy (P = .017)
and focal areas of hyperpigmentation (P = .002) than fellow eyes with dark irides. For eyes with
a choroidal neovascular membrane (CNVM), those with light irides had significantly lower
visual acuities (P = .006) and larger scars (P = .006) than eyes with dark irides. In addition,
extent of disease in the eye with a CNVM was positively correlated with extent of disease in the
fellow eye for most comparisons.
Conclusions. These observations suggest that light iris pigmentation is associated with more
extensive retinal disease in patients with unilateral neovascular AMD. Furthermore, in such
patients, those with worse disease in the eye with a CNVM may tend to have more extensive
atrophic disease in the fellow eye. Invest Ophthalmol Vis Sci. 1 904;35:2734-2740.
1 here is a continuing interest in identifying risk factors for age-related macular degeneration (AMD) that
might provide insight into its cause and, possibly,
treatment. Iris pigmentation has been considered by
several studies. There have been reports of an increased prevalence of AMD among patients with light
irides in the United States1^ and of a higher incidence
in white patients in England than in black patients in
Africa.3 AMD resulting in blindness was found to be
limited to white patients in East Baltimore.4 In contrast, other studies have found no difference based on
iris pigmentation in the frequency of affected ScanFrom the lierman-Cund hibomtory for the Study nj Hi'tiiuil Degeneration*,,
Harvard Medical School, Massachusetts /:y and Far Infirmary, lioston.
Massachusetts.
Sul/pnrted by grants jrom the. National E\e Institute (FY08398). the Rettnitis
Pigmentosa Foundation Fighting Blindness, and the Massachusetts Lions F.ye
Research Fund, Inc.
Submitted for publicatum January 13. 1993; revised November 2, 1993; accepted
January 6. I99'f.
Proprietary interest category: N.
Reprint reprints: Dr. Michael A. Sandberg, lierman-Ciiinil Ijiborators.
Massachusetts E\e and Far Injirmary. 2't3 Charles Street, Boston, MA 02114.
2734
dinavians5 or in the prevalence of AMD among whites
and blacks in the United States.6'' A recent case-control study8 reported that the prevalence of neovascular
AMD, which accounts for the majority of cases of
AMD with legal blindness,9 was unrelated to the degree of iris pigmentation. The reasons for these inconsistent findings with respect to iris pigmentation and
prevalence of AMD are no! clear. For a different perspective, we studied whether light iris pigmentation
was associated with more extensive disease among patients already affected, because a risk factor might
alter severity of expression.
MATERIALS AND METHODS
Patients
We reviewed visual and funduscopic findings of 132
white patients (age range, 58 to 89 years; 73 women
and 59 men) with unilateral neovascular AMD
enrolled in a prospective study of risk factors for the
development of neovascularization in the fellow eye.
Imcsliuaiivr (.)|j|iili.ilniulu^ & Visual Sacnci:, May l'J'.M. Vol. :»:"), No. I)
(;<>|>yi iglu (0 A.ssoi iation loi R c v a n h in Vision and Opliihalniulo^-
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933404/ on 05/14/2017
Iris Pigmentation and Macular Degeneration
2735
Except where noted, findings are from the baseline
visit. Patients were referred primarily from retinal specialists in New England; a few patients were self-referred. Eligibility criteria were the presence of a choroidal neovascular membrane (CNVM) in the affected
eye, a corrected Snellen visual acuity > 20/(50 and sufficiently clear media to allow adequate evaluation of
the fundus in the fellow eye, macular drusen in both
eyes, and no sign of other retinal disease. Eligible patients also had to be able to fixate within the central 2°
with their fellow eye as determined by ophthalmoscopy.10 To our knowledge, iris color was not considered in the referral of any patient. This report includes
all white patients meeting the eligibility criteria who
were enrolled in our prospective study; to rule out a
possible confounding effect of racial background on
iris pigmentation, the black patient who was enrolled
was not included in this presentation.
The diagnosis of unilateral neovascular AMD was
confirmed by ocular examination, fundus photographs, fluorescein angiograms, and previous records.
Based on fundus photographs of fellow eyes, 83% of
the patients had at least some soft, conlluent drusen or
pigmentary changes in the macula. Extent of involvement of the macula by drusen was estimated to have a
median value of 4% and a mean value of 7%; these are
similar to the mean value (4.8%) reported recently for
the fellow eyes of patients with unilateral geographicatrophy or exudative macular degeneration in a population study." All patients gave informed consent after
the nature and possible consequences of the study had
been explained. The procedures followed the tenets of
the Declaration of Helsinki and were approved by the
internal review boards of the Massachusetts Eye and
Ear Infirmary and Harvard Medical School.
Iris color was determined both by asking the patient and by examiner observation with the aid of a
penlight, before evaluation of visual function, pupillary dilation, and assessment of retinal appearance. In
all but one case, the two judgments agreed; for the
exception, the judgment of the examiner was coded.
Iris pigmentation was then categorized dichotomously12 as light (blue, green, or hazel) or dark
(brown). Table 1 compares demographic and other
characteristics for these two groups of patients, including some that have been considered as risk factors
for AMD. Eighty-nine patients were coded as having
light irides, and 43 patients were coded as having dark
irides. The patients with light irides were not significantly different from those with dark irides with respect to mean age, percentage by sex, mean refractive
error (equivalent sphere), percentage with a positive
family history of AMD, percentage with cardiovascular
disease, or percentage with high blood pressure. The
two groups were also not significantly different with
respect to dietary intake of nutrients (including supplements) based on a standardized food-frequency questionnaire12 that converts the weekly intake of 61 food
items into 55 nutrients; this questionnaire was completed after the baseline visit by 98% of available patients. No median nutrient value differed significantly
between the two groups of patients (see Table 1 for
examples); this was true even when analyses were done
by sex (not shown).
TABLE l.
Demographic and Other Characteristics Versus
Iris Pigmentation
Characteristic
Light Iris
Dark Iris
Number
Age (mean ± SE)
Males
Refractive error (mean ± SE)
+ Family Hx of AMD
Cardiovascular disease
High blood pressure
Dietary protein (median)
Dietary animal fai (median)
Dietary vegetable fat (median)
Dietary cholesterol (median)
Dietary carolenoids (median)
Dietary vitamin A (median)
Dietary vitamin E (median)
Dietary calcium (median)
Dietary zinc (median)
89 Patients
74.7 ± 0.7 yrs
40%
+0.82 ± 0.21 D
25%
27%
44%
74.C>8 g
31.20 g
30.01 g
252.4 g
12570 IU
10800 IU
11.17 mgH
790.0 ing
13.35 mg
43 Patients
73.8 ± 1.1 yrs
53%
+ 1.33 ±0.31 D
19%
21%
37%
74.18 g
28.51 g
29.48 g
248.7 g
9524 IU
15810 IU
9.0 mgH
082.4 mg
12.41 mg
* Student's /-test.
+ Chi-square lesi.
I Mann-Whitney test.
'i Alplui-locopherol equivalents.
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933404/ on 05/14/2017
P Value
0.485*
0.158f
0.170*
0.434f
0.453f
0.470f
0.035+
0.080+
0.410+
0.984+
0.430+
0.531 +
0.934+
0.201 +
0.324+
2736
Investigative Ophthalmology & Visual Science, May 1994, Vol. 35, No. 6
Assessment of Visual Function
Best-corrected visual acuity was measured at 18 feet
on a projected Snellen chart with a background luminance of ~ 120 cd/m2 or, for many eyes with a CN VM,
by finger counting. The latter was converted to decimal acuity based on distance from the patient. For
fellow eyes, we remeasured visual acuity with a chart
used by the Early Treatment Diabetic Retinopathy
Study (ETDRS)1' projected onto the screen at the
same background luminance. This chart contains five
letters of comparable difficulty on each line; letters on
adjacent lines differ in size by 0.1 log unit. Visual
acuity was quantified as the number of letters identified (45 letters = 20/20), which provided a more continuous distribution than that for Snellen acuity. For
eight patients whose ETDRS chart acuities were not
available, we estimated them from the linear regression of ETDRS chart acuity on Snellen acuity based on
the remaining patients (R = .87, P < .001).
We measured visual field impairment of fellow
eyes by a method of letter recognition perimetry and
by the macular threshold test of the Humphrey Field
Analyzer. The number of field defects on each of these
tests was previously reported to be strongly correlated
with the extent of retinal pigment epithelial (RPE) atrophy among patients with AMD.10 Letter recognition
perimetry, formerly referred to as computerized
acuity mapping of the macula or CAMM,H involved
identification of upper-case black letters flashed consecutively on a computer screen (Macintosh, Apple
Computer, Cupertino, CA) in random alphabetical
order and location within the central 10° X 10°. Each
letter wasflashedto one of 100 locations distributed in
a grid pattern of 1 ° intervals and to fixation provided
by a cross in the center. The screen had a luminance of
100 cd/m2 and was viewed at a distance of 1 foot with
the appropriate optical correction; the fellow eye was
occluded. While fixating on the cross, patients clicked
the "mouse" to cause the cross to disappear and to
flash a letter for 20 msec. Patients attempted to name
the letter, and their answers were then entered at the
keyboard by the examiner. This returned the fixation
cross to the screen, and the patient repeated the procedure. Letter size increased with increasing distance
from the cross according to a formula10 that compensated for the normal decline in recognition of flashed
letters of fixed size with increasing eccentricity. Locations where letters were missed were retested after the
initial run to help eliminate errors due to lapses of
attention or fixation or due to misidentification of ambiguous letters. The computer then generated an
error count. Because the upper (95%) limit of normal
based on 31 volunteers (age range, 63 to 84 years) was
four errors, this value was subtracted from each patient's score and negative values were converted to
zero. The number of resulting errors was then considered as the number of field defects.
The macular threshold test of the Humphrey Field
Analyzer (model 611, Allergan Humphrey, San Leandro, CA) was administered to a consecutive subset of
patients (n = 97) on a visit after baseline with the standard stimulus (Goldmann size 111, 200 msec duration)
and background (10 cd/m2) and with the pupil monitor on; the fellow eye was patched, and the room lights
were dimmed. Sensitivities in decibels were determined for the 16 locations within a 4° X 6° rectangular matrix. Locations with reduced sensitivity were
identified based on estimated normal age-corrected
95% confidence limits provided by the manufacturer
and were considered as field defects.
Assessment of Retinal Appearance
In addition to ihe ophthalmoscopic examination, macular appearance was evaluated from color slides projected onto a screen. Photographs were of sufficient
quality to allow detailed evaluation in 129 (98%) of the
fellow eyes and in 121 (92%) of the eyes with a CNVM.
For the fellow eyes, we estimated the percentage of
macula with RPE atrophy within a circle drawn on a
plastic overlay that subtended 15°. The atrophy
ranged from patchy hypopigmentation to large, geographic, sharply demarcated, almost totally while
areas of RPE absence. We also counted the number of
focal areas of RPE hyperpigmentation within the circle. These fellow eyes were also rated as low risk or
high risk for neovascularization based on their clinical
appearance with respect to the extent of soft, confluent drusen and RPE pigmentary changes, in accordance with criteria set forth in previous studies;1'"1"
this dichotomous variable, like iris pigmentation, was
tested as a risk factor for predicting macular loss in the
eyes with a CNVM. For the eyes with a CNVM, rated in
sessions separate from those used to rate fellow eyes,
we estimated the equivalent diameter of macular area
in disc diameters affected by scarring, hemorrhage,
and exudation (hereafter referred to as scar diameter).
Macular appearance of a given eye was rated by observers masked to that patient's iris pigmentation, visual findings, and retinal appearance of the other eye.
Statistical Analyses
Because the frequency distributions were skewed for
most outcome measures, we used the distribution-free
Mann-Whitney test for differences in mean rank to
assess the relationship of each outcome measure to iris
pigmentation. We also used this test to relate scar diameter of eyes with a CNVM to risk for neovascularization of fellow eyes. A multiple regression model was
then used to determine the relationship of scar diameter, converted to logarithms to approximate a normal
distribution, to both iris pigmentation and risk for neo-
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933404/ on 05/14/2017
Iris Pigmentation and Macular Degeneration
2737
vascularization of the fellow eye. With this model, the
effect of each factor on scar diameter was evaluated
while controlling for the effect of the other factor. The
association of these two factors, iris pigmentation and
risk for neovascularization of the fellow eye, was determined by chi-square analysis. Lastly, we assessed
the interocular correlation between each pair of outcome measures by the Spearman rank correlation.
Analyses were done with JMP, version 2.0.4 (SAS Institute, Cary, NC) on a Macintosh computer.
RESULTS
Fellow Eye Correlations With Iris Pigmentation
Figure ] illustrates plots of the variation by iris pigmentation for the 10th, 25th, 50th (median), 75th,
and 90th percentiles of the cumulative frequency distribution of scores for each outcome measure in fellow
eyes. For each measure, the median and other percen-
tiles for eyes with light irides tend to correspond to
worse scores compared to those for eyes with dark
irides. Eyes with light irides had marginally lower visual acuities (P = .156) but significantly more visual
field defects by letter recognition perimetry (P= .011)
and by the Humphrey Field Analyzer (P = .043), more
RPE atrophy (P = .017), and more focal areas of RPE
hyperpigmentation (P = .002) than eyes with dark irides.
CNVM Eye Correlations With Iris
Pigmentation and the Fellow Eye's Risk for
Neovascularization
For the eyes with a CNVM, those with light, irides had
significantly lower visual acuities (P = .006) and larger
scars (P = .006) than those with dark irides (Fig 2).
Irrespective of iris pigmentation, scar diameter was
also found to be significantly larger for patients whose
fellow eye was rated at high risk for neovascularization
(n = 63) than for patients whose fellow eye was rated at
Percentile Distributions for Vision and Macular Appearance
in Fellow Eyes versus Iris Pigmentation
1.3"
1.0"
18
I
161412"
0.8-
I
10"
0.6"
0.50.4"
0.3
25th
8"
6"
11
42-
•10th
Percentile
P-.156
Light
o•2
Dark
Light
Dark
~
1
1
i
1
LightP-.043
Dark
16
141210"
8"
6-
I
42-
o-2Light
Dark
P-.002
Light
Dark
FIGURE l. Pcrccniile plots of decimal visual acuity, the number of visualfielddefects based on
letter recognition perimetry (maximum = 97) and on the macular threshold test of the
Humphrey Field Analyzer (maximum = 16), the percentage area of RPE atrophy, and the
number of focal areas of RPE hyperpigmentation in fellow eyes versus iris pigmentation. P
values are for the Mann-Whitney test. Visual acuities are decimal equivalents based on the
number of letters read on a projected ETDRS chart.
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933404/ on 05/14/2017
2738
Investigative Ophthalmology 8c Visual Science, May 1994, Vol. 35, No. 6
low risk (n = 58) (P < .001). Scar diameter remained
significantly higher in eyes with light irides (P < .001)
and in patients whose fellow eye was considered at
high risk for neovascularization (P = .002) when the
effect of each factor on scar diameter was determined
while controlling for the effect of the other factor (Fig
3). Because iris pigmentation and the risk for neovascularization were independent of each other (P =
0.962) and of comparable strength in predicting scar
diameter (Ra = 0.085 and 0.071, respectively), their
combination provided twice the predictive power (R2
= 0.154) of either alone. On average, patients with
light irides and a fellow eye judged to be at high risk
for neovascularization had a scar diameter in the opposite eye that was 44% larger than that of patients
with dark irides and a fellow eye judged to be at low
risk for neovascularization.
Interocular Correlations
Table 2 presents the interocular correlation for each
pair of outcome measures (dependent variables). The
correlations were significant and in the expected direction for visual acuity of the eye with a CNVM versus
number of errors on letter recognition perimetry and
on the macular threshold test by testing the fellow eye.
The correlations were significant and in the expected
direction also for scar diameter in the eye with a
CNVM versus visual acuity, number of errors on letter
recognition perimetry, number of errors on the macular threshold test, and percent area of atrophy determined for the fellow eye.
DISCUSSION
This study shows that light iris pigmentation was associated with more extensive macular disease than dark
iris pigmentation in both eyes of patients with unilat-
Percentile Distributions for Vision and Macular Appearance
in Eyes with a CNVM versus Iris Pigmentation
Light
Dark
Light
FIGURE 2. Percentile plots of decimal visual acuity and scar
diameter in disc diameters in eyes with a CNVM versus iris
pigmentation, /-'values are tor the Mann-Whitney test.
Scar Diameter in Eyes with a CNVM versus Iris Pigmentation
and versus Funduscopic Risk for CNVM in Fellow Eye
.40!
Light
Dark
Iris Pigmentation
Low
High
Risk for CNVM (Fellow Eye)
FIGURE 3. Mean ± SE for scar diameter in log,,, disc diameters in eyes with a CNVM versus iris pigmentation and versus
the risk for neovascularization based on macular appearance
in the fellow eye. Hotted values for the effect of iris pigmentation on scar diameter have been adjusted for the effect of
risk for neovascularization in the fellow eye on scar diameter, and vice versa. P values arc for Student's /-tests.
eral neovascular AMD. Fellow eyes of patients with
lighter pigmented irides tended to have more visual
field impairment by letter recognition perimetry and
by the macular threshold test of the Humphrey Field
Analyzer and more RPE atrophy and focal areas of
RPE hyperpigmentation. On the other hand, visual
acuity did not vary significantly with iris pigmentation;
this could reflect, in part, our eligibility constraint that
patients have acuities > 20/60 in their fellow eyes. For
eyes with a CNVM, those with lighter irides tended to
have worse visual acuities and larger scars than those
with darker irides.
These findings may help to explain the inconsistent results of studies that compared AMD prevalence
to iris pigmentation. Because referrals tend to be
biased toward patients with symptoms, the increased
prevalence of light irides among patients with AMD in
some case—control studies1"' may be due, in pan, to
the tendency for patients with light irides to have more
extensive disease (and symptoms) than patients with
dark irides. Similarly, the finding that recruited patients with light irides were generally younger than
those with dark irides2 might be explained if the
former became symptomatic at an earlier age. In contrast, referrals based on the presence of neovascular
disease, as in a study that failed to find an association
between prevalence of AMD and iris pigmentation,"
would probably not depend upon symptoms deriving
from milder, atrophic changes. Although our results
do not bear upon the prevalence of AMD, they raise
the possibility that severity of disease, as it relates to
iris pigmentation, may have been a confounding factor
in some previous studies.
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933404/ on 05/14/2017
Iris Pigmentation and Macular Degeneration
TABLE 2.
2739
Interocular Correlations for Extent of Disease
Neovascular Eye
Fellow Eye
Visual acuity (decimal)
Visual acuity (letters)
Letter recognition perimetry (errors)
Macular threshold test (errors)
RPE atrophy (% area)
Focal hyperpigmentation (areas)
Scar diameter (DD)
R*
Visual acuity (letters)
Letter recognition periinetry (errors)
Macular threshold test (errors)
RPE atrophy (% area)
Focal hyperpiginentaiion (areas)
.11
-.22
-.31
.03
-.09
-.22
.20
.30
.21
.11
P Value
.221
.015
.003
.758
.361
.016
.001
.004
.021
.236
* Spearman rank coiTelaiion cocHicient.
We also observed a positive relationship, independent of iris pigmentation, between scar diameter in the
eye with a CNVM and the presence of more extensive
soft, confluent drusen and pigmentary changes in the
fellow eye, factors thought to increase the risk for neovascularization. 1:>~10 Because iris pigmentation and the
funduscopic risk for neovascularization of the fellow
eye were uncorrelated and, therefore, independent
predictors, their combination in our regression model
provided greater power than did either alone. Several
studies have shown that drusen have little, if any,
direct bearing upon visual loss in eyes with dry
AMD, 1020 " 2 ' for which reason the present study did
not consider the amount of drusen as an outcome
measure. Instead, the clinical significance of extensive
soft, confluent drusen may be to increase not only the
likelihood, but also the consequence, of the neovascular process. When this risk factor is combined with
light iris pigmentation in a given patient, the prognosis
for neovascularization may be particularly grave.
The link between iris pigmentation as an independent variable and extent of disease in both eyes of our
patients as dependent variables naturally raised the
probability of a correlation between the dependent
variables as well, which was confirmed for most outcome measures. Thus, among our patients with unilateral neovascular AMD, those with worse disease in the
eye with a CNVM tended to have worse atrophic disease in the fellow eye. This interocular correlation may
rellect a cause and effect relationship. For example,
scar size due to a CNVM may depend upon the degree
of preexisting atrophy, as inferred from the retinal
appearance of the fellow eye. Alternatively, the extent
of disease in both eyes may be governed, in part, by
some common factor (systemic, environmental, genetic). Whichever was true, it would in no way affect
the validity of the finding that iris pigmentation helps
to predict visual and retinal outcome in both eyes.
A physiological basis for this proposed association
remains speculative. It is thought by some that in-
creased ocular melanin, which is positively correlated
with increased iris pigmentation, 2 may protect the retina through better absorption of light, and scavenging
of free radicals produced by photo-oxidation.2--"1 Perhaps corroborative is the epidemiologic evidence that,
long-term outdoor light exposure promotes the occurrence of severe macular degeneration 25 ' 21 ' and that,
wearing eyeglasses, sunglasses, or a hat may protect
against the appearance of retinal signs of AM D.2t> However, another study found that the prevalence of neovascular disease was unrelated to history of sunlight
exposure. 8 Alternatively, iris pigmentation in our sample of patients may be a surrogate for some ocular,
medical, or ethnic risk factor for AMD that modifies
disease expression. Review of our patients' refractive
errors, family histories, medical histories, and diets,
however, revealed no significant variation based upon
iris pigmentation. Nevertheless, the possibility that
disease severity is related to another factor correlated
with iris pigmentation needs to be further addressed.
Key Words
macular degeneration, neovasculari/.aiion, iris pigmentation, visual acuity, visual field
References
1. Hyman LG, Lilienfeld AM, Ferris FL 111, Fine SL.
Senile macular degeneration: A case-control study.
Am J Epidemiol. J 983; 118:213-227.
2. Weitcr JJ, Delori FC, Wing GL, Fitch KA. Relationship of senile macular degeneration to ocular pigmentation. Am J Ophthalmol. 1985; 99:185-187.
3. Gregor Z, Jofte I. Senile macular changes in black
African, lir j Ophthalmol. 1978;62:547-550.
4. Sommcr A, Tielsch JM, Kat/.J, et al. Racial differences
in the cause-specific prevalence of blindness in East
Baltimore. N EnglJ Mad. 1991;325:1412-1417.
5. Vinding T. Pigmentation of the eye and hair in relation to age-related macular degeneration. An epidemiological study of 1000 aged individuals. Ada Ophthalmol. (Copenh). 1990; 68:53-58.
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933404/ on 05/14/2017
2740
Investigative Ophthalmology & Visual Science, May 1994, Vol. 35, No. 6
6. Klein BE, Klein R. Cataracts and macular degeneration in older Americans. Arch Ophthalmol. 1982;
100:571-573.
7. Goldberg J, Flowerdew G, Smith E, Brody JA, Tso
MO. Factors associated with agc-relatcd macular degeneration. An analysis of data from the first National
Health and Nutrition Examination Survey. Am J Epidemiol. 1988; 128:700-710.
8. The Eye Disease Case-Control Study Group. Risk factors for neovascular age-related macular degeneration. Arch Ophthalmol. 1992; 110:1701-1708.
9. Ferris FL III, Fine SL, Hyinan LG. Age-related macular degeneration and blindness due to neovascular
maculopathy. Arch Ophthalmol. 1984; 102:1640-1642.
10. Tolentino MJ, Miller S, Gaudio AR, Sandberg MA.
Visual field deficits in early age-related macular degeneration. Vision Res. 1994;34:409-413.
11. Klein R, Klein BEK, Union KLP. Prevalence of agerelated maculopathy: The Beaver Dam Eye Study. Ophthalmology. 1992; 99:933-943.
12. Willett WC, Sampson L, Stampfer MJ, et al. Rcproducibility and validity of a semiquantitative food frequency questionnaire. Am J Epidemiol. 1985; 122:5165.
13. Ferris FL III, Kassoff A. Bresnick GH, Bailey I. New
visual acuity charts for clinical research. AmJ Ophthalmol. 1982;94:91-96.
14. Weiner A, Sandberg MA, Gaudio AR, Kini MM, Berson EL. Hydroxychloroquine retinopathy. AmJ Ophthalmol. 1991;! 12:528-534.
15. Gregor Z, Bird AC, Chisholm IH. Senile disciform
macular degeneration in the second eye. BrJ Ophthalmol. 1977;61:141-147.
16. Strahlman ER, Fine SL, Hillis A. The second eye of
patients with senile macular degeneration. Arch Ophthalmol. 1983; 101:1191-1193.
17. Smiddy WE, Fine SL. Prognosis of patients with bilat-
18.
19.
20.
21.
22.
23.
24.
25.
26.
eral macular drusen. Ophthalmology. 1984;91:271277.
Bressler NM, Bressler SB, Seddon JM, Gragoudas ES,
Jacobson LP. Drusen characteristics in patients with
exudative versus non-exudative age-related macular
degeneration. Retina. 1988;8:109-114.
Bressler SB, Maguire MG, Bressler NM, Fine SL, the
Macular Photocoagulation Study Group. Relationship
of drusen and abnormalities of the retinal pigment
epithelium to the prognosis of neovascular macular
degeneration. Arch Ophthalmol. 1990; 108:14421447.
Sunness JS, Massof RW, Johnson MA, Finkelstein D,
Fine SL. Peripheral retinal function in age-relaied
macular degeneration. Arch Ophthalmol. 1985;
103:811-816.
Eisner A, Fleming SA, Klein ML, Mauldin WM. Sensitivities in older eyes with good acuity: eyes whose fellow eye has exudative disease. Invest Ophthalmol Vis
Sci. 1987; 28:1832-1837.
Sunness JS, Johnson MA, Massof RW, Marcus S. Retinal sensitivity over drusen and nondrusen areas: A
study using fundus perimetry. Arch Ophthalmol.
1988:106:1081-1084.
Eisner A, Stoumbos VD, Klein ML, Fleming SA. Relations between fundus appearance and function: Eyes
whose fellow eye has exudative age-related macular
degeneration. Invest Ophthalmol Vis Sci. 1991; 32:820.
Young RW. Solar radiation and age-related macular
degeneration. Sun Ophthalmol. 1988;32:252-269.
Taylor HR, West S, Mufioz B, Rosenthal FS, Bressler
SB, Bressler NM. The long-term effects of visible light
on the eye. Arch Ophthalmol. 1992; 1 10:99-104.
Cruickshanks KJ, Klein R, Klein BEK. Sunlight and
age-related macular degeneration. The Beaver Dam
Eye Study. Arch Ophthalmol. 1993; 111:514-518.
Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933404/ on 05/14/2017