To explore the correlation between visual acuity (VA) and color vision and to establish a guide for the diagnosis of the cause of visual loss based on this correlation.
Retrospective comparative evaluation of a diagnostic test.
A total of 259 patients with visual impairment caused by 1 of 4 possible disease categories were included. Patients were divided into 4 groups according to the etiology of visual loss: 1) optic neuropathies, 2) macular diseases, 3) media opacities, and 4) amblyopia. The best-corrected VA was established and a standard Ishihara 15 color plates was tested and correlated to the VA in every group separately. Correlation between the VA and the color vision along the different etiologies was evaluated. Frequency of each combination of color vision and VA in every disease category was established.
VA is correlated with color vision in all 4 disease categories. For the same degree of VA loss, patients with optic neuropathy are most likely and patients with amblyopia are the least expected to have a significant color vision loss. Patients with optic neuropathy had considerably worse average color vision (6.7/15) compared to patients in the other 3 disease categories: 11.1/15 (macular diseases), 13.2/15 (media opacities), and 13.4/15 (amblyopia).
Diseases of the optic nerve affect color vision earlier and more profoundly than other diseases. When the cause of visual loss is uncertain, the correlation between the severity of color vision and VA loss can imply the possible etiology of the visual loss.
Occasionally, the ophthalmologist faces a patient in whom the cause of visual loss is uncertain. In such a situation, color vision is generally evaluated since it is assumed that optic nerve disease will affect color vision earlier and more profoundly than any other disease. A significant color vision loss with relatively preserved visual acuity (VA) is regarded as an indication that the origin of visual loss is an optic neuropathy. Surprisingly, this assumption has never been proven. Data about color vision loss in relation to a specific disease can be found in the ophthalmologic literature, but the effect of different classes of ocular disease on color vision has not been evaluated, particularly when diseases with related etiologies are grouped together. Therefore, a comprehensive evaluation of the effect of different categories of ocular diseases on the extent of color vision loss is needed. Such data may be essential when the degree of color vision loss is used as a diagnostic clue.
In this study, we grouped ocular diseases that affect vision into 4 categories: optic neuropathies, macular diseases, media opacities, and amblyopia. Our goal was to answer the following 3 questions: Is there a correlation between VA loss and color vision loss in each category? If so, does this correlation differ between the 4 disease categories? Will the correlational patterns enable the cause of visual loss to be predicted based on the combination of VA and color vision loss?
Two hundred forty-nine patients with visual impairment from different causes were enrolled in this study. Patients who were diagnosed with a specific cause for their visual loss were selected. They were placed into 1 of 4 categories based on the cause of the visual impairment: 1) optic neuropathy, 2) macular disease, 3) media opacity, or 4) amblyopia. We excluded all patients with a visual impairment that could be attributed to more than 1 etiology, or that could be related to other than the 4 categories of visual loss. Male subjects with a severe symmetric bilateral color vision defect were excluded because of the high probability of congenital dyschromatopsia. Patients with glaucoma were also excluded, as VA loss usually only occurs in very advanced disease. When both eyes of a patient met the inclusion criteria to be included in this study, we selected the eye with VA that could give us a better distribution of patients along the spectrum of VA or else the right eye was selected.
Optic neuropathies were diagnosed by a neuro-ophthalmologist (Y.A.). Included in this group were patients with compressing tumors, optic neuritis, and anterior ischemic optic neuropathy. Macular diseases were diagnosed by a retinal specialist and included patients with cystoid macular edema, diabetic macular edema, macular epiretinal membranes, age-related macular degeneration, myopic maculopathies, and macular holes. Media opacities included patients with cataract, corneal, or vitreous opacities. Amblyopia was diagnosed by a patient’s self-report of a unilateral “lazy eye” since early childhood, and only when an amblyogenic factor was established by the examiner (Y.A.).
All patients were tested for both VA and color vision. VA was recorded with the best refractive correction, and then with a pinhole. Patients with VA less than 20/1200 were excluded.
The Ishihara color vision test (Ishihara Test 14 Plate Book, Kanehara Shuppan Co Ltd, Tokyo, Japan) was used to assess color vision. The test was performed according to instructions. Since we presented each subject with15 plates (including the first control plate), the test result was expressed as the number of correctly identified plates on a scale of 1 to 15. If a patient recognized only the control plate, it was recorded as 1/15. If a patient recognized only 1 of a paired number, it was recorded as a half-correct answer. For presbyopic patients we used the near correction for the test. Color vision and VA testing were consistently performed in the same room under identical testing conditions and by the same person (Y.A.). The examinations were performed at the acute or the chronic stage of the disease.
For the statistical analysis, Snellen visual acuities were transformed into decimal fractions. In order to differentiate between the levels of VA and color vision, patient results were divided into 3 levels of VA loss (mild, 20/20-20/100; moderate, <20/100-20/240; and severe, <20/240-20/1200) and 2 levels of color vision (preserved, 10-15; and significant color vision loss, 0-9.5) based on the number of correctly read plates.
To determine if there was any difference between the 4 disease categories, univariate analysis of variance was performed with disease type as the independent variable and VA, color vision, gender, and age as dependent variables. Because of variance differences, a nonparametric test (Kruskal-Wallis test) was performed.
Statistical analysis was performed with SPSS (SPSS, Inc, Chicago, Illinois, USA) programs.
Two hundred and forty-nine patients (107 men and 142 women), with an average age of 58.6 ± 18.8 years (range 16-96 years), were enrolled in this study over a period of 30 months. Demographic details of the patients and mean VA and color vision are presented in Table 1 . Of the entire sample, 87 patients (34.9%) had optic neuropathy, 47 (18.9%) had macular disease, 66 (26.5%) had media opacity, and 49 (19.7%) had amblyopia. Patients differed significantly in age (F (3,225) = 11.75, P < .05), as those with optic neuropathy and amblyopia were younger than patients with macular disease and media opacity.
|Disease||No. of Patients (%)||Average Age (Years)||Males N (%)||Mean Visual Acuity (± SD)||Mean Color Vision (± SD)|
|Optic neuropathy||87 (34.9)||55||35 (40.0)||0.36(.3)=20/55||6.7/15 (5.7)|
|Macular disease||47 (18.9)||64.1||22 (46.8)||0.26(.2)=20/75||11.1/15 (5.2)|
|Media opacity||66 (26.5)||67.5||28 (42.4)||0.35(.2)=20/60||13.2/15 (2.8)|
|Amblyopia||49 (19.7)||49||22 (44.9)||0.27(.2)=20/75||13.4/15 (3.7)|
Significant differences between diseases were found in color vision (F(3,248) = 33.811, P < .001). The average color vision (6.73/15) of patients with optic neuropathy was significantly lower ( P < .01) compared to patients with macular disease (11.1/15), media opacity (13.2/15), or amblyopia (13.4/15) even though their average VA of 0.36 (20/55) was superior to patients of the other groups: 0.26 (20/75), 0.35 (20/60), and 0.27 (20/75), respectively.
The figure displays the relationships between VA and color vision in each disease category. Color vision was correlated to VA in each of the 4 disease categories ( P < .01). The scatter plots demonstrate that patients with optic neuropathy who had mild VA loss (0.2 [20/100] or better) were scattered quite randomly between the levels of color vision—some had preserved color vision while others had significant color vision loss—whereas in the other 3 categories all patients with mild VA loss had preserved color vision. Most patients with moderate VA (0.2 [20/100] to 0.08 [20/240]) in the optic neuropathy and maculopathy groups showed significant color vision loss, while the media opacity and amblyopia groups had virtually intact color vision. At the level of poor VA (less than 20/240 [0.08]), all patients with optic neuropathy had poor color vision while in the other 3 disease categories color vision was either preserved or significantly affected.
The distribution of patients with significant loss of color vision for each level of VA loss for the 4 disease categories is presented in Table 2 . In order to study the probability of having a certain disease for each combination of VA and color vision levels, χ 2 test was performed for each level, with disease as the dependent variable. The proportion of having a certain disease in different subgroups of VA and color vision was statistically significant (χ 2 (33) = 116.45, P < .001). In patients with good VA, significant color vision loss (fewer than 10 of 15 correctly identified plates) was present in 18 (38%) of patients with optic neuropathy but in none of the patients in the other 3 disease categories. In patients with moderate VA loss, significant color vision loss was present in 24 (83%) optic neuropathy patients, 5 (45%) maculopathy patients, and 5 (38%) media opacity patients, but in only 1 (8%) amblyopia patient. In patients with poor VA (less than 0.08 [20/240]), significant color vision loss was present in all optic neuropathy patients but also in most of the patients from the other 3 groups.