The Effect of Early Cataracts on Glare and Contrast Sensitivity A Pilot
Study
Ingrid Adamsons, MD, MPH; Gary S. Rubin, PhD; Susan Vitale, MHS; Hugh R. Taylor, MD; Walter J. Stark, MD \s=b\ To
establish the effect of cataracts on
glare and contrast sensitivity, we graded type and amount of lens opacity in 110 subjects who underwent two glare tests (Brightness Acuity Tester and Berkeley glare test) and two contrast sensitivity tests (a sine-wave test and Pelli-Robson chart). Twenty-seven subjects (25%) had clear lenses (mean visual acuity of 20/20) and 83 subjects (75%) had early lens opacities (mean visual acuity of 20/40) in otherwise normal eyes. Multiple regression techniques were used to control for the effects of age and visual acuity. Glare test scores were significantly lower for nearly all patients with lens opacities than for patients with clear lenses and were the lowest for patients with lenses with posterior subcapsular opacity. Contrast sensitivity scores were lower for all patients with lens opacities than for patients with clear lenses at high frequencies only; all lens opacity groups scored similarly with each other. These results indicate reduced visual function among patients with cataracts whose visual acuity is only
minimally impaired. {Arch Ophthalmol. 1992;110:1081-1086)
Cataract
extraction is
performed
in
^ individuals with cataract who have
decreased visual function. Some of these patients have little decrease in their visual acuity but are believed by their ophthalmologist to have signifi¬ cant glare disability and/or contrast sensitivity loss. These aspects of visual function are often quantified by a vari-
Accepted for publication March 18,
1992. From the Dana Center for International OphAdamsons and thalmology (Drs Vitale) and the Lions Vision Center (Dr Rubin), Wilmer Eye Institute, and the Wilmer Eye Institute (Dr Stark), The Johns Hopkins Medical Institutes, Baltimore, Md, and the Department of Ophthalmology, University of Melbourne, Australia (Dr
Taylor). Reprint requests to the Dana Center, Wilmer 117, The Johns Hopkins Hospital, 600 N Wolfe St, Baltimore, MD 21287-9019 (Dr Adamsons).
ety of
tests
designed
to
SUBJECTS, MATERIALS,
objectively
evaluate glare and contrast sensitivity. These test scores are then used to sup¬ port the decision to perform surgery. However, the association of glare dis¬ ability and contrast sensitivity loss with the presence and type of cataracts has not been well characterized. Neverthe¬ less, a large number of cataract opera¬ tions are probably performed each year primarily to treat glare and contrast
sensitivity problems. Lens opacities scatter stray light en¬ tering the eye and may result in dis¬ abling glare and diminished contrast sensitivity.13 Since visual acuity mea¬
visual function under conditions of maximum contrast (black letters against white) and in the absence of stray illumination, ideal conditions un¬ der which most visual tasks are not performed, visual acuity may overesti¬ mate day-to-day visual functioning. De¬ vices purporting to measure glare and contrast sensitivity have been used to evaluate patients with cataract, and the literature seems to support an associa¬ tion of glare disability and contrast sensitivity loss in individuals with cata¬ racts, even when visual acuity is sures
good.1·2·4"6
However, the studies to date are somewhat limited methodologically. Lens opacities have not been consis¬ tently evaluated and documented1·2·6-15 even though individuals with different types and amount of cataract may be differently affected by glare and vary in their sensitivity to contrast.16"18 Addi¬ tionally, a number of different tests of glare and contrast sensitivity that have variable reliability and sensitivity have been used.1·2·6·8"15·19"25 This makes the test results difficult to compare. Finally, there are several methods of test ad¬ ministration, and the method can sig¬ nificantly affect results.2430 Thus, evi¬ dence of the association of cataract with glare and contrast sensitivity loss needs to be carefully established.
Downloaded From: http://archopht.jamanetwork.com/ by a Yale University User on 05/15/2015
AND
METHODS
Study Population We recruited subjects from those seen in a general ophthalmology practice and from a practice specializing in anterior segment dis¬ ease; appropriate institutional review board approval had been obtained and all subjects gave informed consent. Subjects with cata¬ ract were recruited whether or not they were planning cataract surgery. All subjects un¬ derwent measurement of visual acuity with their best manifest refraction, slit-lamp ex¬ amination, and lens and fundus examination following pupillary dilation. Lenses were evaluated with the Wilmer Eye Institute, Baltimore, Md, system for the clinical grad¬ ing of lens opacities (see below).3133 Subjects with cataract were required to have best cor¬ rected visual acuity of 20/80 or better in their study eye; controls were required to have best corrected visual acuity of20/30 or better. Except for the presence of lens opacities, no ocular disease was present in any of the sub¬ jects. Each subject contributed only one eye to the study. By choosing modest degrees of lens opaciflcation as cutoff points for our cases, we included those patients for whom surgery would not be obviously indicated on the basis of visual acuity alone. Each subject wore the refractive correc¬ tion obtained by his or her physician at the last office visit throughout testing. All test¬ ing was performed using a strict forcedchoice testing procedure to obtain results of the greatest sensitivity and reliability.30 All tests
were
performed monocularly.
Visual Acuity Measurement Visual acuity was tested with the charts and procedures developed by the Early Treatment Diabetic Retinopathy Study,34 which conform to the standards for acuity testing proposed by the National Academy of Science-National Research Council Com¬ mittee on Vision.35 The visual acuity chart was transilluminated with the chart illumi¬ nator (Lighthouse Low Vision Products, Long Island City, NY) light box, which maintained chart luminance at 130 candelas per square meter. The subject was required to name the letters on the chart even if they appeared illegible until all letters on a row were named incorrectly. Visual acuity was scored as the total number of letters read correctly, and converted to LogMAR (loga-
Table
—1
1.—Opacity Types Posterior
Posterior
Opacity Type* (%) of patlentsf Age, y Visual acuity Nuclear opacity, U Cortical opacity, total fraction No.
of involved lens area Posterior subcapsular opacity, mm2
Cortical Nuclear Cortical and Subcapsular Subcapsular and Nuclear Opacity Nuclear Opacity Clear Lens Opacity Opacity Opacity 10(9.1) 19(17.3) 8(7.3) 30(27.3) 11(10) 27(24.5) 72.4(62-87) 56.4(26-85) 51.6(26-69) 70.2(64-85) 68.9(43-87) 65.1(53-77) 20/20(20/13-20/30) 20/35(20/20-20/80) 20/35(20/20-20/53) 20/45(20/27-20/80) 20/40(20/20-20/67) 20/40(20/20-20/80) 3.1(2.5-3.9) 0.8(0-1.8) 1.7(1-1.9) 3.1(2-4.0) 3.2(2-4.0) 0.9(0-1.9)
0(0-1/8) 0(0)
0
33.5(1-100)
*Unless otherwise specified, data are mean (range). tTwo subjects with posterior subcapsular and cortical
rithm of minimum angle resolvable) units, according to the method recommended by Bailey and Lovie.36 Visual acuity was mea¬ sured at the time of glare and contrast sen¬
sitivity testing.
Contrast
Sensitivity Tests
Pelli-Robson Chart.—The Pelli-Robson chart37 consists of 16 groups of three upper¬ case letters that are of constant size but vary in contrast. The groups decrease in contrast by approximately 0.15 log units, ranging from 90% contrast at the upper left to 0.5% contrast at the lower right. The test was ad¬ ministered like an ordinary visual acuity test. The subject named the letters until two or more errors were made in a group. Con¬ trast threshold was determined by the last group in which at least two of the three let¬ ters were correctly identified. Contrast sen¬ sitivity was recorded as the reciprocal of the contrast threshold value. The chart was viewed from 1 m, a distance at which the let¬ ters are equivalent to 20/640 Snellen (0.36 cycles per degree [cpd]). Cathode-Ray Tube (CRT) Test.—Sinewave gratings were generated with a graph¬ ics board (PC Pattern Generator, Vision Met¬ rics, Berkeley, Calif) and displayed on a CRT11 screen (Joyce Electronics, Joyce Elec¬ tronics Ltd, Cambridge, England), with P4 phosphor. The gratings could be displayed with up to 80% contrast. A two-alternative forced-choice staircase procedure was used to measure contrast thresholds. Vertical sine-wave gratings were rotated under com¬ puter control to 15° clockwise or counter¬ clockwise from vertical. The grating was dis¬ played for 1 second (abrupt onset and offset) and the subject indicated whether the grat¬ ing appeared to be tilted to the left or to the right. Ten practice trials were run at high contrast, and then contrast was set to about three times the expected threshold. Follow¬ ing each correct response, contrast was re¬ duced, and following each incorrect response contrast was increased; contrast threshold was the mean ofthe last eight reversals. Con¬ trast threshold was measured at five spatial frequencies, ie, 1.5, 3, 6, 12, and 18 cpd, which were presented in random order.
Disability Glare Tests Brightness Acuity Tester (BAT).—The BAT25 is a 60-mm-diameter hemisphere with a diffusing surface that is placed in front of
the eye. There is a 12-mm central aperture through which a visual acuity chart is viewed; a shielded light bulb is the glare source and
(O-Ve)
7/l6
(1/4-1/2+)
0(0)
o
(o-y8)
0
0(0)
(0-Ve)
17.6(1-64)
%
(1/4-1/2+)
0(0)
opacities and three subjects with all three opacity types were excluded because of the small sample size.
is located superior to the aperture. The sub¬ ject viewed an Early Treatment Diabetic Retinopathy Study letter chart through the central aperture and named the letters on the chart until the responses were no longer cor¬ rect. This acuity was measured first without any glare illumination to obtain a baseline score, and then with the glare light turned on the medium setting (340 cd/m2). Different versions of the Early Treatment Diabetic Retinopathy Study letter chart were used so that memorization of the letters during the course of the test was avoided. The drop in the number of letters read correctly with glare compared with the baseline was the glare disability score. Berkeley Glare Test.—This device38 con¬ sists of a reduced Bailey-Lovie low-contrast
(10% contrast) letter chart36 that is mounted
on a modified slide-viewing box that has three levels of surround-glare illumination of 3000 cd/m2, 800 cd/m2, and 300 cd/m2. The letters on the chart are arranged in 13 rows of five uppercase letters ranging in size from Snellen equivalent 20/160 to 20/10; the de¬ gree of contrast of the letters remains con¬ stant throughout the chart. The subject was seated 1 m from the chart and named the letters on the chart until the responses were no longer correct. The procedure was first performed without any glare illumination to obtain a baseline score and was repeated with high, medium, and low levels of glare il¬ lumination. Different letter charts were used at the different glare levels so that memori¬ zation of the chart during the course of the test was avoided. The drop in the number of letters read correctly at each glare level compared with the baseline was the glare score for each condition.
Photodocumentation and Evaluation of Nuclear Opacities and Color
opacities were photographed pharmacologically dilated pupil. Standardized photographs33 were taken with a photographic slit-lamp camera (Topcon SL-5D Photoslitlamp Camera, Southern Op¬ tical, Greensboro, NC) using standard set¬ tings. The slit beam was 0.3 mm wide and 9 mm long and was angled 30° to the visual axis; it was focused on the center of the nu¬ cleus. The photographs were read on a light box by a trained masked grader and were graded by comparison with a set of four standard slit-lamp photographs, as de¬ scribed previously.31 Previous work with this system confirmed satisfactory agreement between clinical grading and photograding Nuclear
through
a
Downloaded From: http://archopht.jamanetwork.com/ by a Yale University User on 05/15/2015
and
good intraobserver agreement among graders of the lens photographs.31 For the purposes of this study, nuclear opacity mea¬ sured as 2.0 U or greater was classified as significant lens opacity. Nuclear color was graded by comparison with the standard photograph No. 2 that was used in grading nuclear opacity.33 Grade 0 nuclear color was less yellow-brown than that of the light reflected from the posterior capsule of standard photograph No. 2. Grade 1 was similar to that of photograph No. 2, and grade 2 nuclear color was more yellowbrown that that of the standard photograph. Photodocumentation and Evaluation of Cortical and Posterior Subcapsular
(PSC) Opacities photographed through a pharmacologically dilated pupil. Retroilluminated photographs were taken with a retLenses
were
roillumination camera (Neitz Retroillumination Camera, Southern Optical) using standard settings. The photographs were read on a light box by a trained masked grader. When grading cortical opacities, the photographs were placed on an underlay that divided the photographed lenses into equal, wedge-shaped sixteenths. Cortical spokes and wedges were graded as the total number of Vie fractions of total lens area involved. Single opaque lens fibers were not included. Previous work with this system has con¬ firmed good agreement between clinical and photographic grading of cortical opacities and good intraobserver agreement for pho¬ tograph evaluation.39 When grading PSC opacities, the overall vertical and horizontal dimensions of the photographed opacity were measured in millimeters using a cali¬ brated underlay that was prepared by pho¬ tographing a millimeter scale with the retroillumination camera. For the purposes of this study, cortical opacity measured as greater than 2/ie and any PSC opacity were classified as significant lens opacities. Data Analysis
The data were analyzed with a statistical software package (SAS, Version 6.06, SAS Institute Ine, Cary, NC). Preliminary analy¬ sis consisted of descriptive statistics on age, visual acuity, and amount of lens opacity for each opacity type. When data involving the amount of PSC opacity were analyzed, the square root of the area involved with PSC opacity was used. Multiple regression models were used to adjust for age and visual acuity when examining various cataract types.
Table 2.—Patients Limited in Their Performance No. I
Posterior subcapsular (n=19) Cortical (n=8) Nuclear (n=30) Posterior subcapsular and nuclear Cortical and nuclear (n=10)
Berkeley Test I
Low_Medium_High
0
0
2(11) 3(38)
1(9)
4(21) 1(12) 4(13) 2(18)
1(10)
0
1 (4) 5(26)
0
3(10) (n=11)
the
(%) of Patients Limited by Glare Condition
_Opacity Type_Baseline Clear (n=27)
on
0
2(18) 2(20)
3(11) 4(21) 3(38)
from analysis because of the small sam¬ ple size of these subgroups. Subjects with "cortical and nuclear" lenses were excluded from analysis of the BAT data because missing data resulted in a sam¬ ple that was too small for meaningful
analysis.
Glare Tests
11(37)
2(18) 1(10)
Fig 1.—Berkeley glare test scores dropped with increasing glare for all types of lenses. The drop was most pronounced for lenses with posterior subcapsular (PSC) opacity.
Glare
was
measured with the Berke¬
ley glare test (Fig 1) and the BAT (Fig 2). Multiple regression analysis showed that all glare test results were unaffected by visual acuity (Pa .7) or by patient age (P>.1), except for the high-glare condi¬ tion of the Berkeley test, for which pa¬ tient age was significant (P=.01); the re¬
sults of this one test condition have been adjusted for patient age. Many patients were not able to perform the Berkeley glare test at all because they were unable to read any of the low-contrast letters at the baseline no-glare condition (Table 2). In addition, many patients were limited in their performance because, with glare, they were unable to read any of the let¬ ters. For the individuals who could per¬ form at baseline, test scores declined with increasing glare for all lens types (Fig 1). Patients with pure PSC opacity were the most disabled by glare, followed by those with mixed PSC and nuclear opacities: the amount of PSC opacity was considerably less in the mixed group than that in the group with only PSC opacity (36 vs 16.5 mm2). Patients with cortical, nuclear, and combined cortical and nu¬ clear opacities had scores between those of patients with lenses with PSC opacity and those with clear lenses. Scores for pa¬ tients with clear lenses were statistically higher than those of all other lens groups at all glare conditions except for patients with combined cortical and nuclear opac¬ ities, who scored similarly to patients with clear lenses at low- and high-glare conditions. This one exception may be due to its small sample size. The BAT scores were similar to the medium-glare Berkeley test scores (Fig 2). Patients with clear lenses scored the highest, those with lenses with PSC opacity scored lowest, and the other lens groups scored in between. Contrast
Fig 2.—Brightness Acuity Tester scores at the medium-glare condition dropped for lenses with opacity. The drop was most pronounced for lenses with posterior subcapsular (PSC) opacity. RESULTS
The study population consisted of 110 subjects (Table 1). Subjects with clear lenses made up 24.5% of the population (27 patients), those with only one type of opacity made up 51.8% (57 patients), and those with mixed opacities made up 23.6% of the population (26 patients).
Subjects with nuclear opacity made up single largest group (30 subjects [27.3%]), followed by those with clear lenses (27 subjects [24.5%]), and those with PSC opacity (19 subjects [17.3%]). The subjects in the "posterior subcap¬ the
sular and cortical" group and in the "all three opacities" group were excluded
Downloaded From: http://archopht.jamanetwork.com/ by a Yale University User on 05/15/2015
Sensitivity Tests
Contrast sensitivity was measured with the CRT test (Fig 3) and the PelliRobson chart (Fig 4). Multiple regres¬ sion analysis showed that all contrast sensitivity results were significantly affected by age (P
Study
Ingrid Adamsons, MD, MPH; Gary S. Rubin, PhD; Susan Vitale, MHS; Hugh R. Taylor, MD; Walter J. Stark, MD \s=b\ To
establish the effect of cataracts on
glare and contrast sensitivity, we graded type and amount of lens opacity in 110 subjects who underwent two glare tests (Brightness Acuity Tester and Berkeley glare test) and two contrast sensitivity tests (a sine-wave test and Pelli-Robson chart). Twenty-seven subjects (25%) had clear lenses (mean visual acuity of 20/20) and 83 subjects (75%) had early lens opacities (mean visual acuity of 20/40) in otherwise normal eyes. Multiple regression techniques were used to control for the effects of age and visual acuity. Glare test scores were significantly lower for nearly all patients with lens opacities than for patients with clear lenses and were the lowest for patients with lenses with posterior subcapsular opacity. Contrast sensitivity scores were lower for all patients with lens opacities than for patients with clear lenses at high frequencies only; all lens opacity groups scored similarly with each other. These results indicate reduced visual function among patients with cataracts whose visual acuity is only
minimally impaired. {Arch Ophthalmol. 1992;110:1081-1086)
Cataract
extraction is
performed
in
^ individuals with cataract who have
decreased visual function. Some of these patients have little decrease in their visual acuity but are believed by their ophthalmologist to have signifi¬ cant glare disability and/or contrast sensitivity loss. These aspects of visual function are often quantified by a vari-
Accepted for publication March 18,
1992. From the Dana Center for International OphAdamsons and thalmology (Drs Vitale) and the Lions Vision Center (Dr Rubin), Wilmer Eye Institute, and the Wilmer Eye Institute (Dr Stark), The Johns Hopkins Medical Institutes, Baltimore, Md, and the Department of Ophthalmology, University of Melbourne, Australia (Dr
Taylor). Reprint requests to the Dana Center, Wilmer 117, The Johns Hopkins Hospital, 600 N Wolfe St, Baltimore, MD 21287-9019 (Dr Adamsons).
ety of
tests
designed
to
SUBJECTS, MATERIALS,
objectively
evaluate glare and contrast sensitivity. These test scores are then used to sup¬ port the decision to perform surgery. However, the association of glare dis¬ ability and contrast sensitivity loss with the presence and type of cataracts has not been well characterized. Neverthe¬ less, a large number of cataract opera¬ tions are probably performed each year primarily to treat glare and contrast
sensitivity problems. Lens opacities scatter stray light en¬ tering the eye and may result in dis¬ abling glare and diminished contrast sensitivity.13 Since visual acuity mea¬
visual function under conditions of maximum contrast (black letters against white) and in the absence of stray illumination, ideal conditions un¬ der which most visual tasks are not performed, visual acuity may overesti¬ mate day-to-day visual functioning. De¬ vices purporting to measure glare and contrast sensitivity have been used to evaluate patients with cataract, and the literature seems to support an associa¬ tion of glare disability and contrast sensitivity loss in individuals with cata¬ racts, even when visual acuity is sures
good.1·2·4"6
However, the studies to date are somewhat limited methodologically. Lens opacities have not been consis¬ tently evaluated and documented1·2·6-15 even though individuals with different types and amount of cataract may be differently affected by glare and vary in their sensitivity to contrast.16"18 Addi¬ tionally, a number of different tests of glare and contrast sensitivity that have variable reliability and sensitivity have been used.1·2·6·8"15·19"25 This makes the test results difficult to compare. Finally, there are several methods of test ad¬ ministration, and the method can sig¬ nificantly affect results.2430 Thus, evi¬ dence of the association of cataract with glare and contrast sensitivity loss needs to be carefully established.
Downloaded From: http://archopht.jamanetwork.com/ by a Yale University User on 05/15/2015
AND
METHODS
Study Population We recruited subjects from those seen in a general ophthalmology practice and from a practice specializing in anterior segment dis¬ ease; appropriate institutional review board approval had been obtained and all subjects gave informed consent. Subjects with cata¬ ract were recruited whether or not they were planning cataract surgery. All subjects un¬ derwent measurement of visual acuity with their best manifest refraction, slit-lamp ex¬ amination, and lens and fundus examination following pupillary dilation. Lenses were evaluated with the Wilmer Eye Institute, Baltimore, Md, system for the clinical grad¬ ing of lens opacities (see below).3133 Subjects with cataract were required to have best cor¬ rected visual acuity of 20/80 or better in their study eye; controls were required to have best corrected visual acuity of20/30 or better. Except for the presence of lens opacities, no ocular disease was present in any of the sub¬ jects. Each subject contributed only one eye to the study. By choosing modest degrees of lens opaciflcation as cutoff points for our cases, we included those patients for whom surgery would not be obviously indicated on the basis of visual acuity alone. Each subject wore the refractive correc¬ tion obtained by his or her physician at the last office visit throughout testing. All test¬ ing was performed using a strict forcedchoice testing procedure to obtain results of the greatest sensitivity and reliability.30 All tests
were
performed monocularly.
Visual Acuity Measurement Visual acuity was tested with the charts and procedures developed by the Early Treatment Diabetic Retinopathy Study,34 which conform to the standards for acuity testing proposed by the National Academy of Science-National Research Council Com¬ mittee on Vision.35 The visual acuity chart was transilluminated with the chart illumi¬ nator (Lighthouse Low Vision Products, Long Island City, NY) light box, which maintained chart luminance at 130 candelas per square meter. The subject was required to name the letters on the chart even if they appeared illegible until all letters on a row were named incorrectly. Visual acuity was scored as the total number of letters read correctly, and converted to LogMAR (loga-
Table
—1
1.—Opacity Types Posterior
Posterior
Opacity Type* (%) of patlentsf Age, y Visual acuity Nuclear opacity, U Cortical opacity, total fraction No.
of involved lens area Posterior subcapsular opacity, mm2
Cortical Nuclear Cortical and Subcapsular Subcapsular and Nuclear Opacity Nuclear Opacity Clear Lens Opacity Opacity Opacity 10(9.1) 19(17.3) 8(7.3) 30(27.3) 11(10) 27(24.5) 72.4(62-87) 56.4(26-85) 51.6(26-69) 70.2(64-85) 68.9(43-87) 65.1(53-77) 20/20(20/13-20/30) 20/35(20/20-20/80) 20/35(20/20-20/53) 20/45(20/27-20/80) 20/40(20/20-20/67) 20/40(20/20-20/80) 3.1(2.5-3.9) 0.8(0-1.8) 1.7(1-1.9) 3.1(2-4.0) 3.2(2-4.0) 0.9(0-1.9)
0(0-1/8) 0(0)
0
33.5(1-100)
*Unless otherwise specified, data are mean (range). tTwo subjects with posterior subcapsular and cortical
rithm of minimum angle resolvable) units, according to the method recommended by Bailey and Lovie.36 Visual acuity was mea¬ sured at the time of glare and contrast sen¬
sitivity testing.
Contrast
Sensitivity Tests
Pelli-Robson Chart.—The Pelli-Robson chart37 consists of 16 groups of three upper¬ case letters that are of constant size but vary in contrast. The groups decrease in contrast by approximately 0.15 log units, ranging from 90% contrast at the upper left to 0.5% contrast at the lower right. The test was ad¬ ministered like an ordinary visual acuity test. The subject named the letters until two or more errors were made in a group. Con¬ trast threshold was determined by the last group in which at least two of the three let¬ ters were correctly identified. Contrast sen¬ sitivity was recorded as the reciprocal of the contrast threshold value. The chart was viewed from 1 m, a distance at which the let¬ ters are equivalent to 20/640 Snellen (0.36 cycles per degree [cpd]). Cathode-Ray Tube (CRT) Test.—Sinewave gratings were generated with a graph¬ ics board (PC Pattern Generator, Vision Met¬ rics, Berkeley, Calif) and displayed on a CRT11 screen (Joyce Electronics, Joyce Elec¬ tronics Ltd, Cambridge, England), with P4 phosphor. The gratings could be displayed with up to 80% contrast. A two-alternative forced-choice staircase procedure was used to measure contrast thresholds. Vertical sine-wave gratings were rotated under com¬ puter control to 15° clockwise or counter¬ clockwise from vertical. The grating was dis¬ played for 1 second (abrupt onset and offset) and the subject indicated whether the grat¬ ing appeared to be tilted to the left or to the right. Ten practice trials were run at high contrast, and then contrast was set to about three times the expected threshold. Follow¬ ing each correct response, contrast was re¬ duced, and following each incorrect response contrast was increased; contrast threshold was the mean ofthe last eight reversals. Con¬ trast threshold was measured at five spatial frequencies, ie, 1.5, 3, 6, 12, and 18 cpd, which were presented in random order.
Disability Glare Tests Brightness Acuity Tester (BAT).—The BAT25 is a 60-mm-diameter hemisphere with a diffusing surface that is placed in front of
the eye. There is a 12-mm central aperture through which a visual acuity chart is viewed; a shielded light bulb is the glare source and
(O-Ve)
7/l6
(1/4-1/2+)
0(0)
o
(o-y8)
0
0(0)
(0-Ve)
17.6(1-64)
%
(1/4-1/2+)
0(0)
opacities and three subjects with all three opacity types were excluded because of the small sample size.
is located superior to the aperture. The sub¬ ject viewed an Early Treatment Diabetic Retinopathy Study letter chart through the central aperture and named the letters on the chart until the responses were no longer cor¬ rect. This acuity was measured first without any glare illumination to obtain a baseline score, and then with the glare light turned on the medium setting (340 cd/m2). Different versions of the Early Treatment Diabetic Retinopathy Study letter chart were used so that memorization of the letters during the course of the test was avoided. The drop in the number of letters read correctly with glare compared with the baseline was the glare disability score. Berkeley Glare Test.—This device38 con¬ sists of a reduced Bailey-Lovie low-contrast
(10% contrast) letter chart36 that is mounted
on a modified slide-viewing box that has three levels of surround-glare illumination of 3000 cd/m2, 800 cd/m2, and 300 cd/m2. The letters on the chart are arranged in 13 rows of five uppercase letters ranging in size from Snellen equivalent 20/160 to 20/10; the de¬ gree of contrast of the letters remains con¬ stant throughout the chart. The subject was seated 1 m from the chart and named the letters on the chart until the responses were no longer correct. The procedure was first performed without any glare illumination to obtain a baseline score and was repeated with high, medium, and low levels of glare il¬ lumination. Different letter charts were used at the different glare levels so that memori¬ zation of the chart during the course of the test was avoided. The drop in the number of letters read correctly at each glare level compared with the baseline was the glare score for each condition.
Photodocumentation and Evaluation of Nuclear Opacities and Color
opacities were photographed pharmacologically dilated pupil. Standardized photographs33 were taken with a photographic slit-lamp camera (Topcon SL-5D Photoslitlamp Camera, Southern Op¬ tical, Greensboro, NC) using standard set¬ tings. The slit beam was 0.3 mm wide and 9 mm long and was angled 30° to the visual axis; it was focused on the center of the nu¬ cleus. The photographs were read on a light box by a trained masked grader and were graded by comparison with a set of four standard slit-lamp photographs, as de¬ scribed previously.31 Previous work with this system confirmed satisfactory agreement between clinical grading and photograding Nuclear
through
a
Downloaded From: http://archopht.jamanetwork.com/ by a Yale University User on 05/15/2015
and
good intraobserver agreement among graders of the lens photographs.31 For the purposes of this study, nuclear opacity mea¬ sured as 2.0 U or greater was classified as significant lens opacity. Nuclear color was graded by comparison with the standard photograph No. 2 that was used in grading nuclear opacity.33 Grade 0 nuclear color was less yellow-brown than that of the light reflected from the posterior capsule of standard photograph No. 2. Grade 1 was similar to that of photograph No. 2, and grade 2 nuclear color was more yellowbrown that that of the standard photograph. Photodocumentation and Evaluation of Cortical and Posterior Subcapsular
(PSC) Opacities photographed through a pharmacologically dilated pupil. Retroilluminated photographs were taken with a retLenses
were
roillumination camera (Neitz Retroillumination Camera, Southern Optical) using standard settings. The photographs were read on a light box by a trained masked grader. When grading cortical opacities, the photographs were placed on an underlay that divided the photographed lenses into equal, wedge-shaped sixteenths. Cortical spokes and wedges were graded as the total number of Vie fractions of total lens area involved. Single opaque lens fibers were not included. Previous work with this system has con¬ firmed good agreement between clinical and photographic grading of cortical opacities and good intraobserver agreement for pho¬ tograph evaluation.39 When grading PSC opacities, the overall vertical and horizontal dimensions of the photographed opacity were measured in millimeters using a cali¬ brated underlay that was prepared by pho¬ tographing a millimeter scale with the retroillumination camera. For the purposes of this study, cortical opacity measured as greater than 2/ie and any PSC opacity were classified as significant lens opacities. Data Analysis
The data were analyzed with a statistical software package (SAS, Version 6.06, SAS Institute Ine, Cary, NC). Preliminary analy¬ sis consisted of descriptive statistics on age, visual acuity, and amount of lens opacity for each opacity type. When data involving the amount of PSC opacity were analyzed, the square root of the area involved with PSC opacity was used. Multiple regression models were used to adjust for age and visual acuity when examining various cataract types.
Table 2.—Patients Limited in Their Performance No. I
Posterior subcapsular (n=19) Cortical (n=8) Nuclear (n=30) Posterior subcapsular and nuclear Cortical and nuclear (n=10)
Berkeley Test I
Low_Medium_High
0
0
2(11) 3(38)
1(9)
4(21) 1(12) 4(13) 2(18)
1(10)
0
1 (4) 5(26)
0
3(10) (n=11)
the
(%) of Patients Limited by Glare Condition
_Opacity Type_Baseline Clear (n=27)
on
0
2(18) 2(20)
3(11) 4(21) 3(38)
from analysis because of the small sam¬ ple size of these subgroups. Subjects with "cortical and nuclear" lenses were excluded from analysis of the BAT data because missing data resulted in a sam¬ ple that was too small for meaningful
analysis.
Glare Tests
11(37)
2(18) 1(10)
Fig 1.—Berkeley glare test scores dropped with increasing glare for all types of lenses. The drop was most pronounced for lenses with posterior subcapsular (PSC) opacity.
Glare
was
measured with the Berke¬
ley glare test (Fig 1) and the BAT (Fig 2). Multiple regression analysis showed that all glare test results were unaffected by visual acuity (Pa .7) or by patient age (P>.1), except for the high-glare condi¬ tion of the Berkeley test, for which pa¬ tient age was significant (P=.01); the re¬
sults of this one test condition have been adjusted for patient age. Many patients were not able to perform the Berkeley glare test at all because they were unable to read any of the low-contrast letters at the baseline no-glare condition (Table 2). In addition, many patients were limited in their performance because, with glare, they were unable to read any of the let¬ ters. For the individuals who could per¬ form at baseline, test scores declined with increasing glare for all lens types (Fig 1). Patients with pure PSC opacity were the most disabled by glare, followed by those with mixed PSC and nuclear opacities: the amount of PSC opacity was considerably less in the mixed group than that in the group with only PSC opacity (36 vs 16.5 mm2). Patients with cortical, nuclear, and combined cortical and nu¬ clear opacities had scores between those of patients with lenses with PSC opacity and those with clear lenses. Scores for pa¬ tients with clear lenses were statistically higher than those of all other lens groups at all glare conditions except for patients with combined cortical and nuclear opac¬ ities, who scored similarly to patients with clear lenses at low- and high-glare conditions. This one exception may be due to its small sample size. The BAT scores were similar to the medium-glare Berkeley test scores (Fig 2). Patients with clear lenses scored the highest, those with lenses with PSC opacity scored lowest, and the other lens groups scored in between. Contrast
Fig 2.—Brightness Acuity Tester scores at the medium-glare condition dropped for lenses with opacity. The drop was most pronounced for lenses with posterior subcapsular (PSC) opacity. RESULTS
The study population consisted of 110 subjects (Table 1). Subjects with clear lenses made up 24.5% of the population (27 patients), those with only one type of opacity made up 51.8% (57 patients), and those with mixed opacities made up 23.6% of the population (26 patients).
Subjects with nuclear opacity made up single largest group (30 subjects [27.3%]), followed by those with clear lenses (27 subjects [24.5%]), and those with PSC opacity (19 subjects [17.3%]). The subjects in the "posterior subcap¬ the
sular and cortical" group and in the "all three opacities" group were excluded
Downloaded From: http://archopht.jamanetwork.com/ by a Yale University User on 05/15/2015
Sensitivity Tests
Contrast sensitivity was measured with the CRT test (Fig 3) and the PelliRobson chart (Fig 4). Multiple regres¬ sion analysis showed that all contrast sensitivity results were significantly affected by age (P
