BritishJournal ofOphthalmology 1992; 76: 719-722 719

Contrast sensitivity and glare in cataract using the Br J Ophthalmol: first published as 10.1136/bjo.76.12.719 on 1 December 1992. Downloaded from Pelli-Robson chart

T H Williamson, N P Strong, J Sparrow, R K Aggarwal, R Harrad

Abstract contrast sensitivity and a disposable pen torch as There is a need for a convenient, clinically the glare source. This test is simple and rapid, applicable test of glare disability which can be and could be easily introduced into a busy used in the preoperative evaluation of patients clinical environment. with cataract. In this study, contrast sensitivity (using the Pelli-Robson letter chart), near vision, and visual acuity were compared, with Patients and methods and without the introduction of a glare source Subjects with cataract were recruited from in 70 patients with cataract, 15 with intraocular patients attending the Department of Ophthal- lenses, and 19 controls. A disposable pen torch mology at Addenbrooke's Hospital, Cambridge. was shone at the pupillary margin to induce Patients with high myopia or ocular disease other glare. Contrast sensitivity demonstrated the than cataract were excluded. most marked reduction during glare testing. All subjects were questioned as to the extent of Cortical cataracts were most affected followed their symptoms for near vision, distance vision, by posterior subcapsular opacities. The glare and glare disability and asked to rank these on a disability was significantly less in pseudophakic six point scale. patients and was absent from patients with Snellen visual acuity was tested using a non-cataractous phakic eyes. Glare testing standard back illuminated Snellen chart at 6 m. with a disposable pen torch and a Pelli-Robson The appropriate spectacle correction was worn contrast sensitivity letter chart provides a and the visual acuity was rounded off to the rapid test ofglare disability which can be easily nearest line. incorporated into the clinical appraisal of Jaeger plates were used to test near vision at patients with cataract. 0 3 m with a reading correction. (BrJ Ophthalmol 1992; 76: 719-722) Contrast sensitivity was tested with a Pelli- Robson chart.7 8 This chart consists of letters of constant size arranged in 16 groups ofthree. The A discrepancy is often encountered between the contrast of all the letters in the first of these http://bjo.bmj.com/ visual function measured by Snellen acuity and triplets is 100%. The contrast ofeach subsequent the visual disability that is experienced by triplet is reduced by a factor of 0 707 (0- 15 log patients with cataract. The disparity has become unit). Thus the contrast of the last triplet is more apparent as the reliability, acceptability, 0 56% (2-25 log units below 100%). The chart Tennent Institute of and success rate of cataract extraction and intra- was used at 1 m at a mean luminance of 65 cd/m2 Ophthalmology, Glasgow ocular lens implantation has improved. A as recommended by the suppliers.

GI16NT on September 23, 2021 by guest. Protected copyright. T H Williamson measurement of visual disability by Snellen Glare was introduced with a disposable pen acuity alone is now less than satisfactory and torch which was held at 20 degrees to the visual Department of additional tests ofvisual deficit and postoperative axis at 30 cm from the eye and directed at the Ophthalmology, Royal are At this the pen has a mean Victoria Infirmary, improvement required. pupil. distance torch Newcastle NEI 4LP Contrast sensitivity and glare disability are luminance of 4000 cd/m2. Snellen acuity, Jaeger N P Strong sensitive measurements of visual loss in patients near acuity, and Pelli-Robson contrast sensi- with cataract, particularly those with mild or tivity were tested without and then with the pen Department of Ophthalmiology, moderate lens opacities. Investigators using torch glare. The pen torch was replaced after Leicester Royal computer based monitor systems'-3 or projection every five patients. The pens were tested with a Infirmary, Leicester devices4 have demonstrated a reduction in photometer which confirmed that there was no LEI 5WW contrast in these Paulsson loss of with this amount of usage. J Sparrow sensitivity patients. brightness and Sjostrand3 also tested the effect of glare on The lens opacity of each cataract patient Birmingham and Midland contrast sensitivity by placing a bright light was graded for type (posterior subcapsular, Eye Hospital, Church Street, Birmingham source adjacent to a VDU displaying sinusoidal nuclear, cortical, or mixed) and for severity (on B3 2NS gratings and Abrahamsson and Sjostrand have a three point scale). R K Aggarwal used a TV monitor surrounded by a fluorescent Bristol Eye Hospital, tube for this purpose.5 6 These tests are sensitive Lower Maudlin Street, measures ofvisual function but they are laborious STATISTICAL METHODS Bristol BS1 2LX to perform. The incorporation of these testing For the purpose of statistical analysis the Snellen R Harrad methods into routine practice would require the and Jaeger acuities were converted to equivalent Correspondence to: Dr T H Williamson, Tennent purchase of expensive equipment. If a test is to value of visual angle using the decimal scale in Institute of Ophthalmology, be used regularly in the clinic it is essential that it which 6/6 has the value of 1-0, 6/12 the value 0 5, Western Infirmary, Church Street, Glasgow GIl 6NT. is quick, reproducible, and readily available. In 6/60 the value 0-1, and so on. The effect of glare Accepted for publication this study, we report the results obtained by upon vision for near (J-diff) or distance (VA-diff) 22 July 1992 using a Pelli-Robson letter based chart to test was expressed as the difference between the 720 Williamson, Strong, Sparrow, Aggarwal, Harrad

logarithms of the values obtained with and 2 0 - n = 105 Br J Ophthalmol: first published as 10.1136/bjo.76.12.719 on 1 December 1992. Downloaded from without glare. As the Pelli-Robson chart gives a r=r 075 logarithmic measure of contrast sensitivity the 1 6- a effect of glare upon contrast sensitivity (CS-diff) o was obtained directly by subtraction. Rosner's :LI A multivariate analysis using 1 24 intraclass correlation model was performed. U) This is the most efficient statistical method for a, o

the analysis of studies with a two eye design and / O Cl 0 84 0 continuous dependent variables. The method o 80 0 m 4 0 0 avoids the overestimation of significance levels 0 which occurs if individual eyes are regarded as 0.4-f independent units of observation. The analysis was performed using the Generalised Linear o/oa a a a g Interactive Modelling (GLIM) system of the 0*0 - 8 ° ° Numerical Algorithms Group (Wilkinson House, Jordan Hill Road, Oxford, OX2 8DR). HM CF 6/60 6/24 6/9 In the analyses eyes were classified by lens status Log Snellen acuity into groups and subgroups according to cataract Figure I Contrast sensitivity versus Snellen acuity (both on type (cortical, posterior subcapsular, nuclear, a log scale) prior to surgeryfor all the patients in the and mixed), pseudophakia, and phakic non- study. cataractous controls. When the effect of glare upon contrast sensitivity was compared between the different Results types of cataract there were significant dif- ferences between these groups (p=0 0025). The PATIENT CHARACTERISTICS subtype of cataract producing the greatest loss of The mean age ofthe group with cataract was 72*8 contrast in response- to glare was the cortical, years (SD 9 4, range 44-91); ofthe pseudophakic followed in reducing order by posterior sub- patients was 72-9 years (SD 9 4, range 64-86); capsular, mixed, and nuclear cataract. and of the non-cataractous phakic controls was There were no differences in the glare effect 60-9 years (SD 10-0, range 44-76). upon Snellen acuity between the cataract sub- In the cataract group there were 105 eyes in 70 groups, pseudophakic, or control subjects patients; in the pseudophakic group there were (p=0 34). The glare effect upon Jaeger vision, 15 eyes in 15 patients; and in the control group however, was not uniform across these groups with clear lenses there were 34 eyes in 19 (p=0045). The difference in performance was patients. The 105 eyes with cataract were further accounted for by the non-cataractous phakic divided into subgroups by cataract morphology controls, who were less affected by glare than the

such that there were 17 eyes with cortical cataractous and pseudophakic subjects. http://bjo.bmj.com/ cataract, 28 eyes with posterior subcapsular Compared with the controls the other subgroups cataract, 37 eyes with nuclear cataract, and 23 were affected from most to least in the order: eyes with mixed cataract morphology. posterior subcapsular cataract, mixed cataract, cortical cataract, nuclear cataract, and pseudo- phakia (Table 2). Differences within this rank SNELLEN VISUAL ACUITY. JAEGER VISION AND ordering however were not significant (p=0 13). CONTRAST SENSITIVITY There were marked differences in the glare on September 23, 2021 by guest. Protected copyright. There were no significant differences between effect between the groups and subgroups with the visual acuities without glare in the four regard to contrast sensitivity (p< 10-6). This cataract subgroups (p=009). Without glare, powerful effect was mainly due to differences Snellen acuity correlated well with contrast between the non-cataractous phakic controls and sensitivity (r=0-74), Jaeger acuity correlated the other groups. There was a significant dif- well with contrast sensitivity (r=0-73), and ference between the pseudophakic subjects and Snellen acuity correlated well with Jaeger acuity the non-cataractous controls (0-005), the (r=0-75). Despite these apparently good cor- pseudophakic subjects suffering a more marked relations between the various measures of visual loss ofcontrast sensitivity than the controls. The function, there remains considerable spread pseudophakic subjects experienced slightly less about the regression line between these variables. loss of contrast sensitivity than the subjects with This is exemplified graphically in Figure 1, cataracts (Table 3), though this effect was where it can be seen that for a given Snellen statistically non-significant (p=0 14). acuity there remains quite a wide range of contrast sensitivity. SYMPTOMS Snellen acuity and visual symptoms at distance GLARE EFFECT showed a strong association (p< 10-6) and The pen torch glare effect was regarded as the symptoms of reading difficulty correlated difference in visual function measured with the similarly with Jaeger vision (p<10-6). However glare source off and on. The effect of glare on symptoms of glare difficulty were not signifi- visual function was examined in the groups with cantly associated with the effect of glare upon cataract, pseudophakia, and clear lenses, as well Snellen acuity (p=0 20), Jaeger acuity (pO=071), as in the four cataract subgroups. or contrast sensitivity (p=0-61). Contrast sensitivity andglare in cataract usingthe Pelli-Robson chart 721

Table I The mean distance visual acuityfor each ofthe disability. Different patients may have rated the groups, with and without glare, and the mean effect ofglare. Br J Ophthalmol: first published as 10.1136/bjo.76.12.719 on 1 December 1992. Downloaded from (All data on a linear scale) same degree ofglare disability quite differently. All forms of cataract showed glare disability. Type With glare Without glare Difference In keeping with common clinical experience PSCLO* 0 354 0 374 0-01% posterior subcapsular lens opacities showed Nuclear 0-326 0 340 0-0142 severe glare loss. However the marked extent to Cortical 0-431 0-436 0 0049 Mixed 0-276 0-292 0 0160 which glare also affected the patientswith cortical IOLt 0-606 0-630 0-0240 lens opacities was unexpected. Control 0-887 0-887 0 0000 After cataract extraction and intraocular lens *Posterior subcapsular lens opacity. implantation the glare disability is much reduced tIntraocular lens. but an effect of glare is still demonstrable.'115 In the present study the patients with extra- Table 2 The meanjaeger near acuityfor each ofthe groups, capsular cataract extraction and posterior with and withoutglare, and the mean effect ofglare. (All data chamber lens implantation had a mean reduction on a linear scale) in CS owing to glare of 04152 (Table 3) yet the Type With glare Withoutglare Difference control group was found to have no loss owing to glare at all. This glare disability in pseudophakia PSCLO 0-148 0-200 0-0515 Nuclear 0194 0-222 0-0275 may be attributable to light scattering either by Cortical 0 221 0-283 0-0620 the posterior capsule or by the intraocular lens. Mixed 0 122 0-181 0 0590 IOL 0-306 0-329 0-0227 The advantages of contrast sensitivity and Control 0 354 0-354 0-0000 glare disability measurements in the clinical setting can be illustrated by the following case. A Table 3 The contrast sensitivityfor each ofthe groups, with 23-year-old male, who worked as a labourer, and withoutglare, and the mean effect ofglare presented complaining of difficulty seeing in bright light. He wore sunglasses when he worked Type With glare Withoutglare Difference outdoors but these brought only partial benefit PSCLO 0 701 0 991 0 2896 and he was unable to read a number plate at 25 Nuclear 0 690 0-861 0-1705 Cortical 0 154 0 660 0 4941 yards in daylight and thus unable to drive. He Mixed 0 560 0 777 0-2161 was found to have congenital cataracts, more IOL 0-991 1-143 0-1519 Control 1 453 1 453 0 0000 marked in the right eye, and had already under- gone a right lensectomy 3 years previously. Unfortunately, the visual acuity in this eye Discussion improvedonlyto 6/36withcontact lenscorrection As demonstrated by Figure 1, contrast sensitivity because of the presence of stimulation depri- for a given visual acuity varies quite widely vation amblyopia. The visual acuity in the left between patients. CS loss, therefore measures a eye was 6/6 with appropriate refractive correction different visual disability compared with loss of but the contrast sensitivity on the Pelli-Robson Snellen acuity and can provide the clinician with chart was only 0-80 (normal range 1-65-1-89 for added information concerning the extent of a 20-30-year-old). This became unrecordable in http://bjo.bmj.com/ visual disability.' This is especially the case in the presence of the glare light. A snowstorm patients whose visual acuity remains good despite punctate type of congenital cataract was present significant lens opacities. Contrast sensitivity in the left eye. drops as the severity ofthe cataract increases and Many surgeons would have been reluctant to improves markedly after cataract extraction and carry out a cataract extraction and lens implant lens implantation.9 on this patient in view of the good visual acuity. Patients with cataract often complain about However accurate measurements of the visual on September 23, 2021 by guest. Protected copyright. glare - for example, from bright sunlight or car deficit by contrast sensitivity and glare disability headlights, and may find thisglare more disabling strongly supported the decision to operate. The than a moderate drop in visual acuity. These patient therefore underwent cataract extraction symptoms have been termed glare disability5 10 and lens implantation and subsequently a YAG and are due to increased intraocular light scatter capsulotomy. The postoperative visual acuity and subsequent loss of contrast of the retinal was 6/6 and the contrast sensitivity improved to image. Attempts have been made to provide 1 65 and did not deteriorate with glare. The clinically applicable tests of glare. For instance patient is now able to work without any difficulty Prager et al measured Snellen acuity under office and is planning to take his driving test. light compared with that in sunlight as a measure This study shows that the Pelli-Robson chart of glare" and Maltzman used a pen torch to is a much more sensitive means ofdemonstrating provide glare whilst measuring Snellen visual glare disability than Snellen or Jaeger acuity. acuity. 12 However Snellen acuity is little affected Using the chart and a pen torch it is possible to by loss of contrast so these tests can be expected show in patients with cataract that contrast to be insensitive. Our finding of a much more sensitivity is reduced by glare to a much greater marked effect of glare upon CS than either extent than in normal controls. The Pelli-Robson Snellen acuity or Jaeger near vision suggests that chart is both convenient and quick to use. It is CS is a substantially more sensitive measure of easy for the patient to perform and the observer glare disability than the other two tests. to interpret. In this study it was found to be a We did not find a correlation between the useful clinical test of contrast sensitivity and subjective symptoms of visual glare and the glare disability in patients with cataract. This measured effect ofglare upon contrast sensitivity. test may prove particularly useful in those This may reflect the difficulty of devising a patients whose complaints are out of proportion meaningful scoring system for subjective glare to their Snellen acuity. 722 Williamson, Strong, Sparrow, Aggarwzal, Harrad

1 Hess R, Woo G. Vision through cataracts. Invest Ophthalmol the Pelli-Robson chart. Invest Ophthalmol Vis Sci 1990; 30

Vis Sci 1978; 17: 428-35. (Suppl): 406. Br J Ophthalmol: first published as 10.1136/bjo.76.12.719 on 1 December 1992. Downloaded from 2 Elliot DB, Gilchrist J, Hurst M, Pickwell LD, Sheridan M, 9 Adamsons I, Rubin GS, Abbey H, Stark WJ. Correlation of Weatherill J. The subjective assessment of cataract. visual questionnaire results with glare and contrast test Ophthalmol Physiol Opt 1989; 9: 16-9. results in cataract patients. Invest Ophthalmol Vis Sci 1992; 3 Paulsson LE, Sjostrand J. Contrast sensitivity in the presence 33: 1301. of glare light. Invest Ophthalmol Vis Sci 1980; 19: 10 Hirsch RP, Nadler MP, Miller D. Clinical performance of a 401-6. disability glare tester. Arch Ophthalmol 1984; 102: 1633-6. 4 Hirsch RP, Nadler MP, Miller D, Glare measurements as a 11 Prager TC, Urso RG, Holladay JT, Stewart RH. Glare testing predictor of outdoor vision among cataract patients. Am in cataract patients: instrument evaluation and identification Ophthalmol 1984; 16: 965-8. of sources of methodological error. J7 Cataract Refract Surg 5 Abrahamsson M, Sjostrand J. Impairment of contrast 1989; 15: 149-57. sensitivity function (CSF) as a measure of disability glare. 12 Maltzman BA, Horan C, Rengel A. Penlight test for glare Invest Ophthalmol VisSci 1986; 27: 1131-6. disability of cataracts. Ophthalmic Surg 1988; 19: 356-8. 6 Sjostrand J, Abrahamsson M, Hard AL. Glare disability as a 13 Nadler DJ, Jaffe NS, Clayman HM, Jaffe MS, Luscombe SM. cause of deterioration of vision in cataract patients. Acta Glare disability in eyes with intraocular lenses. Am Ophthalmol Suppl (Kbh) 1986; 182: 103-6. Ophthalmol 1984; 97: 43-7. 7 Pelli DG, Robson JG, Wilkins AJ. The design of a new letter 14 Weatherhill J, Yap M. Contrast sensitivity in pseudophakia chart for measuring contrast sensitivity. Clin Vis Sci 1988; 2: and aphakia. Ophthalmnl Physiol Opt 1986; 6: 297-301. 187-9. 15 Koch DD, Jardeleza TL, Emery JM, Franklin D. Glare 8 Zhang L, Pelli DG, Robson JG. The effects of luminance, following posterior chamber intraocular lens implantation. distance and defocus on contrast sensitivity as measured by CataractRefract Surg 1987; 13: 431-5. http://bjo.bmj.com/ on September 23, 2021 by guest. Protected copyright.